UPLOADER MATCHING METHOD AND DEVICE

PROCEDE ET DISPOSITIF D'APPARIEMENT DE DISPOSITIF DE TELECHARGEMENT AMONT

English Abstract

An uploader rnatching method and device, which belong to the field of computer
information
technologies and are applicable to the field of short videos. The method
comprises: acquiring
released video data of uploaders, determining, according to the released video
data, comprehensive
score values of the uploaders from one or more dimensional feature scores and
preset weights of the
dimensional features, and filtering a target uploader according to the
comprehensive score values of
the uploaders (101); counting released video data of the target uploader
according to a preset tag
rule, and generating corresponding uploader word vectors of one or more video
category tags (102);
acquiring video data played back by a user within a first preset period,
counting the video data
played back by the user according to the preset tag rule, and generating
corresponding user word
vectors of the one or more video category tags (103); performing corresponding
matching between
the corresponding video category tags of the uploader word vectors and the
user word vectors, to
acquire an uploader word vector result that reaches a target degree of
matching with the user, and
determining corresponding uploader information according to the uploader word
vector result (104).

French Abstract

L'invention concerne un procédé et un dispositif d'appariement de dispositif de téléchargement amont, qui appartiennent au domaine des technologies d'informations informatiques et qui peuvent être appliqués au domaine des vidéos courtes. Le procédé consiste à : acquérir des données vidéo libérées de dispositifs de téléchargement amont, déterminer, conformément aux données vidéo libérées, des valeurs de score complètes des dispositifs de téléchargement amont à partir d'un ou de plusieurs scores de caractéristiques dimensionnelles et de poids prédéfinis des caractéristiques dimensionnelles, et filtrer un dispositif de téléchargement amont cible conformément aux valeurs de score complètes des dispositifs de téléchargement amont (101) ; compter les données vidéo libérées du dispositif de téléchargement amont cible conformément à une règle d'étiquette prédéfinie, et générer des vecteurs de mot de dispositif de téléchargement amont correspondants d'une ou de plusieurs étiquettes de catégorie vidéo (102) ; acquérir des données vidéo lues par un utilisateur sur une première période prédéfinie, compter les données vidéo lues par l'utilisateur conformément à la règle d'étiquette prédéfinie, et générer des vecteurs de mot d'utilisateur correspondants de la ou des étiquettes de catégorie vidéo (103) ; effectuer un appariement correspondant entre les étiquettes de catégorie vidéo correspondantes des vecteurs de mot de dispositif de téléchargement amont et des vecteurs de mot d'utilisateur, pour acquérir un résultat de vecteur de mot de dispositif de téléchargement amont qui atteint un degré cible d'appariement avec l'utilisateur, et déterminer des informations de dispositif de téléchargement amont correspondantes conformément au résultat de vecteur de mot de dispositif de téléchargement amont (104).

Claims

Claims:
1. A device for uploader matching comprising:
a calculating module configured to:
obtain a released video data of at least one uploader;
determine at least one comprehensive score value of the at least one
uploader from at least one dimension feature score according to the
released video data; and
screen out at least one target uploader according to the at least one
comprehensive score value of the at least one uploader;
an uploader word vector generating module configured to:
make statistics on the released video data of the at least one target uploader

according to a preset tagging rule; and
generate a corresponding uploader word vector of at least one video
category tag;
a user word vector generating module configured to:
obtain a user played-back video data within a first preset period;
make statistics on the user played-back video data according to the preset
tagging rule; and
generate a corresponding user word vector of the at least one video
category tag; and
a matching module configured to:
3 1
Date Reçue/Date Received 2023-11-21

match the corresponding of the at least one video category tag of the at
least one uploader word vector and a user word vector;
obtain an uploader word vector result that has reached a target matching
degree with the at least one uploader word vector; and
determine corresponding uploader information according to the uploader
word vector result.
2. The device of claim 1 wherein the calculating module further comprises:
a first calculating sub-module configured to calculate at least one score of
one or
more of at least one dimension feature in released video activity scores of
the at least
one uploader, video quality scores of the at least one uploader, and video
verticality
scores of the at least one uploader according to the released video data;
a second calculating sub-module configured to calculate the at least one
comprehensive score value of the at least one uploader according to the one or
more
dimension feature score; and
a screening sub-module configured to select the at least one uploader who rank
above
a threshold to serve as the at least one target uploader according to a
sequence of the
at least one comprehensive score value of the at least one uploader arranged
in a
decreasing order, wherein the threshold is an integer greater than one.
3. The device of claim 2 wherein the first calculating sub-module
configured to:
check an external order information of each target order according to a preset
label
generating rule; and
obtain a checking result of the each target order.
4. The device according to any one of claims 2 to 3 wherein the first
calculating submodule
is further configured to:
32
Date Recue/Date Received 2023-11-21

sort a number of released videos of the at least one uploader within a second
preset
period, and a volume of released videos played back within the second preset
period
respectively in combination with a time decay;
map the sorted number of released videos and the sorted volume of released
videos
played back within the second preset period to a range of [x1,1] and a range
of
[x2,1], wherein x 1 is a first weight index and x2 is a second weight index
each
evaluated as a decimal between 0 and 1;
determine respective weight indices of the number of released videos and the
volume
of released videos played back;
multiply the respective weight indices of the number of released videos with
the
respective weight indices of the volume of released videos played back;
calculate to obtain the released video activity scores of the at least one
uploader;
sort one or more of a number of sharings, a number of praisings, a number of
commentings, a proportion of positive comments, a number of listings as
favorites, a
number of followings and released video playback integrity rates of released
videos
of the at least one uploader within the second preset period respectively
combined
with the time decay;
map the sorted number of sharings, number of praisings, number of commentings,

proportion of positive comments, number of listings as favorites, number of
followings and released video playback integrity rates to a range of [x3,1], a
range of
[x4,1], a range of [x5,1], a range of [x6,1], a range of [x7,1], a range of
[x8,1] and a
range of [x9,1], respectively wherein a x3 is a third weight index, x4 is a
fourth
weight index, x5 is a fifth weight index, x6 is a sixth weight index, x7 is a
seventh
weight index, x8 is an eighth weight index and x9 a ninth weight index;
33
Date Recue/Date Received 2023-11-21

determine respective weight indices of the number of sharings, the number of
praisings, the number of commentings, the proportion of positive comments, the

number of listings as favorites, the number of followings and the released
video
playback integrity rates, wherein a third weight index x3, a fourth weight
index x4, a
fifth weight index x5, a sixth weight index x6, a seventh weight index x7, an
eighth
weight index x8 and a ninth weight index x9 are each evaluated as a decimal
between
0 and 1;
summate and average the respective weight indices of the number of sharings,
the
number of praisings, the number of commentings, the proportion of positive
comments, the number of listings as favorites and the number of followings;
multiply a result of the summating and a result of the averaging with the
respective
weight indices of the released video playback integrity rates;
calculate to obtain the video quality scores of the at least one uploader;
sort category proportions of released videos of the at least one uploader
within the
second preset period in combination with the time decay;
map the sorted category proportions to a range of [x10,1] wherein x10 is a
tenth
weight;
determine respective weight indices of the category proportions, wherein a
tenth
weight index x10 is evaluated as a decimal between 0 and 1;
multiply the respective weight indices of the category proportions; and
calculate to obtain the video verticality scores of the at least one uploader.
5.
The device of claim 4 wherein obtaining the first weight index xl, the second
weight index
x2, the third weight index x3, the fourth weight index x4, the fifth weight
index x5, the
sixth weight index x6, the seventh weight index x7, the eighth weight index
x8, the ninth
weight index x9 and the tenth weight index x10 comprises:
34
Date Recue/Date Received 2023-11-21

taking, respectively, at least one dimension feature score to which the first
weight
index xl, the second weight index x2, the third weight index x3, the fourth
weight
index x4, the fifth weight index x5, the sixth weight index x6, the seventh
weight
index x7, the eighth weight index x8, the ninth weight index x9 and the tenth
weight
index x10 correspond as independent variables;
taking at least one following degree of the at least one uploader after
exposure as
dependent variables; and
employing a RandomForest algorithm and a gradient boosted decision tree (GBDT)

algorithm to calculate the first weight index xl, the second weight index x2,
the third
weight index x3, the fourth weight index x4, the fifth weight index x5, the
sixth
weight index x6, the seventh weight index x7, the eighth weight index x8, the
ninth
weight index x9 and the tenth weight index x10.
6. The device of any one of claims 2 to 5 wherein the second calculating
sub-module is
configured to:
multiply the released video activity scores, the video quality scores and the
video
verticality scores; and
calculate to obtain the at least one comprehensive score value of the at least
one
uploader.
7. The device of any one of claims 1 to 6 wherein the user word vector
generating module is
further configured to:
eliminate one or more of hotspot videos and miss-clicked videos from the user
played-back video data;
count target user tags whose number of videos occupies a proportion that is
not lower
than a preset proportion according to the preset tagging rule wherein the
preset
proportion is greater than one;
Date Recue/Date Received 2023-11-21

calculate the target user tags; and
generate corresponding user word vectors of the target user tags.
8. The device of any one of claims 1 to 7 further comprising a data
recommending module
configured to recommend the uploader information to the user.
9. The device of any one of claims 1 to 8 further comprising the data
recommending module
configured to push a video of the at least one video category tag
corresponding to the
uploader word vector result to the user.
10. The device of any one of claims 1 to 9 wherein one or more of activity
of the at least one
uploader, quality of the at least one uploader, and verticality of the at
least one uploader
are comprehensively considered to facilitate subsequent comprehensive scoring
of a quality
of the at least one uploader.
11. A system for uploader matching comprising:
a calculating module configured to:
obtain a released video data of at least one uploader;
determine at least one comprehensive score value of the at least one
uploader from at least one dimension feature score according to the
released video data; and
screen out at least one target uploader according to the at least one
comprehensive score values of the at least one uploader;
an uploader word vector generating module configured to:
make statistics on the released video data of the at least one target uploader

according to a preset tagging rule; and
36
Date Recue/Date Received 2023-11-21

generate a corresponding uploader word vector of at least one video
category tag;
a user word vector generating module configured to:
obtain a user played-back video data within a first preset period;
make statistics on the user played-back video data according to the preset
tagging rule; and
generate a corresponding user word vector of the at least one video
category tag; and
a matching module configured to:
match the corresponding of the at least one video category tag of the at
least one uploader word vector and a user word vector;
obtain an uploader word vector result that has reached a target matching
degree with the user; and
determine corresponding uploader information according to the uploader
word vector result.
12. The system of claim 11 wherein the calculating module further
comprises:
a first calculating sub-module configured to calculate at least one score of
one or
more of at least one dimension feature in released video activity scores of
the at least
one uploader, video quality scores of the at least one uploader, and video
verticality
scores of the at least one uploader according to the released video data;
a second calculating sub-module configured to calculate the at least one
comprehensive score value of the at least one uploader according to the one or
more
dimension feature score; and
37
Date Recue/Date Received 2023-11-21

a screening sub-module configured to select the at least one uploader who rank
above
a threshold to serve as the at least one target uploader according to a
sequence of the
at least one comprehensive score value of the at least one uploader arranged
in a
decreasing order, wherein the threshold is an integer greater than one.
13. The system of claim 12 wherein the first calculating sub-module
configured to:
check an extemal order information of each target order according to a preset
label
generating rule; and
obtain a checking result of the each target order.
14. The system according to any one of claims 12 to 13 wherein the first
calculating submodule
is further configured to:
sort a number of released videos of the at least one uploader within a second
preset
period, and a volume of released videos played back within the second preset
period
respectively in combination with a time decay;
map the sorted number of released videos and the sorted volume of released
videos
played back within the second preset period to a range of [x1,1] and a range
of
[x2,1], wherein x 1 is a first weight index and x2 is a second weight index
each
evaluated as a decimal between 0 and 1;
determine respective weight indices of the number of released videos and the
volume
of released videos played back;
multiply the respective weight indices of the number of released videos with
the
respective weight indices of the volume of released videos played back;
calculate to obtain the released video activity scores of the at least one
uploader;
38
Date Recue/Date Received 2023-11-21

sort one or more of a number of sharings, a number of praisings, a number of
commentings, a proportion of positive comments, a number of listings as
favorites, a
number of followings and released video playback integrity rates of released
videos
of the at least one uploader within the second preset period respectively
combined
with the time decay;
map the sorted number of sharings, number of praisings, number of commenfings,

proportion of positive comments, number of listings as favorites, number of
followings and released video playback integrity rates to a range of [x3,1], a
range of
[x4,1], a range of [x5,1], a range of [x6,1], a range of [x7,1], a range of
[x8,1] and a
range of [x9,1], respectively wherein a x3 is a third weight index, x4 is a
fourth
weight index, x5 is a fifth weight index, x6 is a sixth weight index, x7 is a
seventh
weight index, x8 is an eighth weight index and x9 a ninth weight index;
determine respective weight indices of the number of sharings, the number of
praisings, the number of commentings, the proportion of positive comments, the

number of listings as favorites, the number of followings and the released
video
playback integrity rates, wherein a third weight index x3, a fourth weight
index x4, a
fifth weight index x5, a sixth weight index x6, a seventh weight index x7, an
eighth
weight index x8 and a ninth weight index x9 are each evaluated as a decimal
between
0 and 1;
summate and average the respective weight indices of the number of sharings,
the
number of praisings, the number of commentings, the proportion of positive
comments, the number of listings as favorites and the number of followings;
multiply a result of the summating and a result of the averaging with the
respective
weight indices of the released video playback integrity rates;
calculate to obtain the video quality scores of the at least one uploader;
39
Date Recue/Date Received 2023-11-21

sort category proportions of released videos of the at least one uploader
within the
second preset period in combination with the time decay;
map the sorted category proportions to a range of [x10,1] wherein x10 is a
tenth
weight;
determine respective weight indices of the category proportions, wherein a
tenth
weight index x10 is evaluated as a decimal between 0 and 1;
multiply the respective weight indices of the category proportions; and
calculate to obtain the video verticality scores of the at least one uploader.
15.
The system of claim 14 wherein obtaining the first weight index xl, the second
weight
index x2, the third weight index x3, the fourth weight index x4, the fifth
weight index x5,
the sixth weight index x6, the seventh weight index x7, the eighth weight
index x8, the
ninth weight index x9 and the tenth weight index x10 comprises:
taking, respectively, at least one dimension feature score to which the first
weight
index xl, the second weight index x2, the third weight index x3, the fourth
weight
index x4, the fifth weight index x5, the sixth weight index x6, the seventh
weight
index x7, the eighth weight index x8, the ninth weight index x9 and the tenth
weight
index x10 correspond as independent variables;
taking at least one following degree of the at least one uploader after
exposure as
dependent variables; and
employing a RandomForest algorithm and a GBDT algorithm to calculate the first

weight index xl, the second weight index x2, the third weight index x3, the
fourth
weight index x4, the fifth weight index x5, the sixth weight index x6, the
seventh
weight index x7, the eighth weight index x8, the ninth weight index x9 and the
tenth
weight index x10.
Date Reçue/Date Received 2023-11-21

16. The system of any one of claims 12 to 15 wherein the second calculating
sub-module is
configured to:
multiply the released video activity scores, the video quality scores and the
video
verticality scores; and
calculate to obtain the at least one comprehensive score value of the at least
one
uploader.
17. The system of any one of claims 11 to 16 wherein the user word vector
generating module
is further configured to:
eliminate one or more of hotspot videos and miss-clicked videos from the user
played-back video data;
count target user tags whose number of videos occupies a proportion that is
not lower
than a preset proportion according to the preset tagging rule wherein a preset

proportion is greater than one;
calculate the target user tags; and
generate corresponding user word vectors of the target user tags.
18. The system of any one of claims 11 to 17 further comprising a data
recommending module
configured to recommend the uploader information to the user.
19. The system of any one of claims 11 to 18 further comprising the data
recommending
module configured to push a video of the at least one video category tag
corresponding to
the uploader word vector result to the user.
20. The system of any one of claims 11 to 19 wherein one or more of
activity of the at least
one uploader, quality of the at least one uploader, and verticality of the at
least one uploader
are comprehensively considered to facilitate subsequent comprehensive scoring
of a quality
of the at least one uploader.
41
Date Recue/Date Received 2023-11-21

21. A method for uploader matching comprising:
obtaining a released video data of at least one uploader;
determining at least one comprehensive score value of the at least one
uploader from
at least one dimension feature score according to the released video data; and
screening out at least one target uploader according to the at least one
comprehensive
score value of the at least one uploader;
making statistics on the released video data of the at least one target
uploader
according to a preset tagging rule; and
generating a corresponding uploader word vector of at least one video category
tag;
obtaining a user played-back video data within a first preset period;
making statistics on the user played-back video data according to the preset
tagging
rule; and
generating a corresponding user word vector of the at least one video category
tag;
and
matching the corresponding of the at least one video category tag of the at
least one
uploader word vector and a user word vector;
obtaining an uploader word vector result that has reached a target matching
degree
with the user; and
determining corresponding uploader information according to the uploader word
vector result.
22. The method of claim 21 further comprising:
42
Date Recue/Date Received 2023-11-21

calculating at least one score of one or more of at least one dimension
feature in
released video activity scores of the at least one uploader, video quality
scores of the
at least one uploader, and video verticality scores of the at least one
uploader
according to the released video data;
calculating the at least one comprehensive score value of the at least one
uploader
according to the one or more dimension feature score; and
selecting the at least one uploader who rank above a threshold to serve as the
at least
one target uploader according to a sequence of the at least one comprehensive
score
value of the at least one uploader arranged in a decreasing order, wherein the

threshold is an integer greater than one.
23. The method of claim 22 further comprising:
checking an external order information of each target order according to a
preset label
generating rule; and
obtaining a checking result of the each target order.
24. The method according to any one of claims 22 to 23 further comprising:
sorting a number of released videos of the at least one uploader within a
second
preset period, and a volume of released videos played back within the second
preset
period respectively in combination with a time decay;
mapping the sorted number of released videos and the sorted volume of released

videos played back within the second preset period to a range of [x1,1] and a
range
of [x2,1], wherein xl is a first weight index and x2 is a second weight index
each
evaluated as a decimal between 0 and 1;
determining respective weight indices of the number of released videos and the

volume of released videos played back;
43
Date Recue/Date Received 2023-11-21

multiplying the respective weight indices of the number of released videos
with the
respective weight indices of the volume of released videos played back;
calculating to obtain the released video activity scores of the at least one
uploader;
sorting one or more of a number of sharings, a number of praisings, a number
of
commentings, a proportion of positive comments, a number of listings as
favorites, a
number of followings and released video playback integrity rates of released
videos
of the at least one uploader within the second preset period respectively
combined
with the time decay;
mapping the sorted number of sharings, number of praisings, number of
commentings, proportion of positive comments, number of listings as favorites,

number of followings and released video playback integrity rates to a range of
[x3,1],
a range of [x4,11, a range of [x5,1], a range of [x6,1], a range of [x7,1], a
range of
[x8,1] and a range of [x9,1], respectively wherein a x3 is a third weight
index, x4 is a
fourth weight index, x5 is a fifth weight index, x6 is a sixth weight index,
x7 is a
seventh weight index, x8 is an eighth weight index and x9 a ninth weight
index;
determining respective weight indices of the number of sharings, the number of

praisings, the number of commentings, the proportion of positive comments, the

number of listings as favorites, the number of followings and the released
video
playback integiity rates, wherein a third weight index x3, a fourth weight
index x4, a
fifth weight index x5, a sixth weight index x6, a seventh weight index x7, an
eighth
weight index x8 and a ninth weight index x9 are each evaluated as a decimal
between
0 and 1;
summating and averaging the respective weight indices of the number of
sharings,
the number of praisings, the number of commentings, the proportion of positive

comments, the number of listings as favorites and the number of followings;
44
Date Recue/Date Received 2023-11-21

multiplying a result of the summating and a result of the averaging with the
respective weight indices of the released video playback integrity rates;
calculating to obtain the video quality scores of the at least one uploader;
sorting category proportions of released videos of the at least one uploader
within the
second preset period in combination with the time decay;
mapping the sorted category proportions to a range of [x10,1] wherein x10 is a
tenth
weight;
determining respective weight indices of the category proportions, wherein a
tenth
weight index x10 is evaluated as a decimal between 0 and 1;
multiplying the respective weight indices of the category proportions; and
calculating to obtain the video verticality scores of the at least one
uploader.
25.
The method of claim 24 wherein obtaining the first weight index xl, the second
weight
index x2, the third weight index x3, the fourth weight index x4, the fifth
weight index x5,
the sixth weight index x6, the seventh weight index x7, the eighth weight
index x8, the
ninth weight index x9 and the tenth weight index x10 comprises:
taking, respectively, at least one dimension feature score to which the first
weight
index xl, the second weight index x2, the third weight index x3, the fourth
weight
index x4, the fifth weight index x5, the sixth weight index x6, the seventh
weight
index x7, the eighth weight index x8, the ninth weight index x9 and the tenth
weight
index x10 correspond as independent variables;
taking at least one following degree of the at least one uploader after
exposure as
dependent variables; and
Date Recue/Date Received 2023-11-21

employing a RandomForest algorithm and a GBDT algorithm to calculate the first

weight index xl, the second weight index x2, the third weight index x3, the
fourth
weight index x4, the fifth weight index x5, the sixth weight index x6, the
seventh
weight index x7, the eighth weight index x8, the ninth weight index x9 and the
tenth
weight index x10.
26. The method of any one of claims 22 to 25 further comprising:
multiplying the released video activity scores, the video quality scores and
the video
verticality scores; and
calculating to obtain the at least one comprehensive score value of the at
least one
uploader.
27. The method of any one of claims 21 to 26 further comprising:
eliminating one or more of hotspot videos and miss-clicked videos from the
user
played-back video data;
counting target user tags whose number of videos occupies a proportion that is
not
lower than a preset proportion according to the preset tagging rule wherein a
preset
proportion is greater than one;
calculating the target user tags; and
generating corresponding user word vectors of the target user tags.
28. The method of any one of claims 21 to 27 further comprising
recommending the uploader
information to the user.
29. The method of any one of claims 21 to 28 further comprising pushing a
video of the at least
one video category tag corresponding to the uploader word vector result to the
user.
46
Date Recue/Date Received 2023-11-21

30. The method of any one of claims 21 to 29 wherein one or more of the
activity of the at least
one uploader, quality of the at least one uploader, and verticality of the at
least one uploader
are comprehensively considered to facilitate subsequent comprehensive scoring
of a quality
of the at least one uploader.
31. A computer equipment for uploader matching comprising a computer
readable physical
memory and a processor communicatively connected to the memory wherein the
processor
is configured to execute a computer-executable instructions stored on the
memory and
wherein the processor when executing the computer-executable instructions is
configured
to:
obtain a released video data of at least one uploader
determine at least one comprehensive score value of the at least one uploader
from at
least one dimension feature score according to the released video data; and
screen out at least one target uploader according to the at least one
comprehensive
score value of the at least one uploader;
make statistics on the released video data of the at least one target uploader
according
to a preset tagging rule; and
generate a corresponding uploader word vector of at least one video category
tag;
obtain a user played-back video data within a first preset period;
make statistics on the user played-back video data according to the preset
tagging
rule; and
generate a corresponding user word vector of the at least one video category
tag; and
match the corresponding of the at least one video category tag of the at least
one
uploader word vector and a user word vector;
47
Date Recue/Date Received 2023-11-21

obtain an uploader word vector result that has reached a target matching
degree with
the user; and
determine corresponding uploader information according to the uploader word
vector
result.
32. The computer equipment of claim 31 wherein the processor is further
configured to:
calculate at least one score of one or more of at least one dimension feature
in
released video activity scores of the at least one uploader, video quality
scores of the
at least one uploader, and video verticality scores of the at least one
uploader
according to the released video data;
calculate the at least one comprehensive score value of the at least one
uploader
according to the one or more dimension feature score; and
select the at least one uploader who rank above a threshold to serve as the at
least one
target uploader according to a sequence of the at least one comprehensive
score value
of the at least one uploader arranged in a decreasing order, wherein the
threshold is
an integer greater than one.
33. The computer equipment of claim 32 wherein the processor is further
configured to:
check an external order information of each target order according to a preset
label
generating rule; and
obtain a checking result of the each target order.
34. The computer equipment according to any one of claims 32 to 33 wherein
the processor is
further configured to:
sort a number of released videos of the at least one uploader within a second
preset
period, and a volume of released videos played back within the second preset
period
respectively in combination with a time decay;
48
Date Recue/Date Received 2023-11-21

map the sorted number of released videos and the sorted volume of released
videos
played back within the second preset period to a range of [x1,1] and a range
of
[x2,1], wherein x I is a first weight index and x2 is a second weight index
each
evaluated as a decimal between 0 and 1;
determine respective weight indices of the number of released videos and the
volume
of released videos played back;
multiply the respective weight indices of the number of released videos with
the
respective weight indices of the volume of released videos played back;
calculate to obtain the released video activity scores of the at least one
uploader;
sort one or more of a number of sharings, a number of praisings, a number of
commentings, a proportion of positive comments, a number of listings as
favorites, a
number of followings and released video playback integrity rates of released
videos
of the at least one uploader within the second preset period respectively
combined
with the time decay;
map the sorted number of sharings, number of praisings, number of commentings,

proportion of positive comments, number of listings as favorites, number of
followings and released video playback integrity rates to a range of [x3,1], a
range of
[x4,1], a range of [x5,1], a range of [x6,1], a range of [x7,1], a range of
[x8,1] and a
range of [x9,1], respectively wherein a x3 is a third weight index, x4 is a
fourth
weight index, x5 is a fifth weight index, x6 is a sixth weight index, x7 is a
seventh
weight index, x8 is an eighth weight index and x9 a ninth weight index;
49
Date Recue/Date Received 2023-11-21

determine respective weight indices of the number of sharings, the number of
praisings, the number of commentings, the proportion of positive comments, the

number of listings as favorites, the number of followings and the released
video
playback integrity rates, wherein a third weight index x3, a fourth weight
index x4, a
fifth weight index x5, a sixth weight index x6, a seventh weight index x7, an
eighth
weight index x8 and a ninth weight index x9 are each evaluated as a decimal
between
0 and 1;
summate and average the respective weight indices of the number of sharings,
the
number of praisings, the number of commentings, the proportion of positive
comments, the number of listings as favorites and the number of followings;
multiply a result of the summating and a result of the averaging with the
respective
weight indices of the released video playback integrity rates;
calculate to obtain the video quality scores of the at least one uploader;
sort category proportions of released videos of the at least one uploader
within the
second preset period in combination with the time decay;
map the sorted category proportions to a range of [x10,1] wherein x10 is a
tenth
weight;
determine respective weight indices of the category proportions, wherein a
tenth
weight index x10 is evaluated as a decimal between 0 and 1;
multiply the respective weight indices of the category proportions; and
calculate to obtain the video verticality scores of the at least one uploader.
35.
The computer equipment of claim 34 wherein obtaining the first weight index x
1, the
second weight index x2, the third weight index x3, the fourth weight index x4,
the fifth
weight index x5, the sixth weight index x6, the seventh weight index x7, the
eighth weight
index x8, the ninth weight index x9 and the tenth weight index x10 comprises:
Date Recue/Date Received 2023-11-21

taking, respectively, at least one dimension feature score to which the first
weight
index xl, the second weight index x2, the third weight index x3, the fourth
weight
index x4, the fifth weight index x5, the sixth weight index x6, the seventh
weight
index x7, the eighth weight index x8, the ninth weight index x9 and the tenth
weight
index x10 correspond as independent variables;
taking at least one following degree of the at least one uploader after
exposure as
dependent variables; and
employing a RandomForest algorithm and a GBDT algorithm to calculate the first

weight index xl, the second weight index x2, the third weight index x3, the
fourth
weight index x4, the fifth weight index x5, the sixth weight index x6, the
seventh
weight index x7, the eighth weight index x8, the ninth weight index x9 and the
tenth
weight index x10.
36. The computer equipment of any one of claims 32 to 35 wherein the
processor is further
configured to:
multiply the released video activity scores, the video quality scores and the
video
verticality scores; and
calculate to obtain the at least one comprehensive score value of the at least
one
uploader.
37. The computer equipment of any one of claims 31 to 36 wherein the
processor is further
configured to:
eliminate one or more of hotspot videos and miss-clicked videos from the user
played-back video data;
count target user tags whose number of videos occupies a proportion that is
not lower
than a preset proportion according to the preset tagging rule wherein a preset

proportion is greater than one;
51
Date Recue/Date Received 2023-11-21

calculate the target user tags; and
generate corresponding user word vectors of the target user tags.
38. The computer equipment of any one of claims 31 to 37 wherein the
processor is further
configured to recommend the uploader information to the user.
39. The computer equipment of any one of claims 31 to 38 wherein the
processor is further
configured to push a video of the at least one video category tag
corresponding to the
uploader word vector result to the user.
40. The computer equipment of any one of claims 31 to 39 wherein one or
more of the activity
of the at least one uploader, quality of the at least one uploader, and
verticality of the at
least one uploader are comprehensively considered to facilitate subsequent
comprehensive
scoring of a quality of the at least one uploader.
41. A computer readable physical memory having stored upon it a computer-
executable
instructions when executed by a computer configured to:
obtain a released video data of at least one uploader;
determine at least one comprehensive score value of the at least one uploader
from at
least one dimension feature score according to the released video data; and
screen out at least one target uploader according to the at least one
comprehensive
score value of the at least one uploader;
make statistics on the released video data of the at least one target uploader
according
to a preset tagging rule; and
generate a corresponding uploader word vector of at least one video category
tag;
obtain a user played-back video data within a first preset period;
52
Date Recue/Date Received 2023-11-21

make statistics on the user played-back video data according to the preset
tagging
rule; and
generate a corresponding user word vector of the at least one video category
tag; and
match the corresponding of the at least one video category tag of the at least
one
uploader word vector and a user word vector;
obtain an uploader word vector result that has reached a target matching
degree with
the user; and
determine corresponding uploader information according to the uploader word
vector
result.
42. The memory of claim 41 wherein the computer is further configured to:
calculate at least one score of one or more of at least one dimension feature
in
released video activity scores of the at least one uploader, video quality
scores of the
at least one uploader, and video verticality scores of the at least one
uploader
according to the released video data;
calculate the at least one comprehensive score value of the at least one
uploader
according to the one or more dimension feature score; and
select the at least one uploader who rank above a threshold to serve as the at
least one
target uploader according to a sequence of the at least one comprehensive
score value
of the at least one uploader arranged in a decreasing order, wherein the
threshold is
an integer greater than one.
43. The memory of claim 42 wherein the computer is further configured to:
check an external order information of each target order according to a preset
label
generating rule; and
53
Date Recue/Date Received 2023-11-21

obtain a checking result of the each target order.
44.
The memory according to any one of claims 42 to 43 wherein the processor is
further
configured to:
sort a number of released videos of the at least one uploader within a second
preset
period, and a volume of released videos played back within the second preset
period
respectively in combination with a time decay;
map the sorted number of released videos and the sorted volume of released
videos
played back within the second preset period to a range of [x1,1] and a range
of
[x2,1], wherein x I is a first weight index and x2 is a second weight index
each
evaluated as a decimal between 0 and 1;
determine respective weight indices of the number of released videos and the
volume
of released videos played back;
multiply the respective weight indices of the number of released videos with
the
respective weight indices of the volume of released videos played back;
calculate to obtain the released video activity scores of the at least one
uploader;
sort one or more of a number of sharings, a number of praisings, a number of
commentings, a proportion of positive comments, a number of listings as
favorites, a
number of followings and released video playback integrity rates of released
videos
of the at least one uploader within the second preset period respectively
combined
with the time decay;
54
Date Recue/Date Received 2023-11-21

map the sorted number of sharings, number of praisings, number of commentings,

proportion of positive comments, number of listings as favorites, number of
followings and released video playback integrity rates to a range of [x3,1], a
range of
[x4,1], a range of [x5,1], a range of [x6,1], a range of [x7,1], a range of
[x8,1] and a
range of [x9,1], respectively wherein a x3 is a third weight index, x4 is a
fourth
weight index, x5 is a fifth weight index, x6 is a sixth weight index, x7 is a
seventh
weight index, x8 is an eighth weight index and x9 a ninth weight index;
determine respective weight indices of the number of sharings, the number of
praisings, the number of commentings, the proportion of positive comments, the

number of listings as favorites, the number of followings and the released
video
playback integrity rates, wherein a third weight index x3, a fourth weight
index x4, a
fifth weight index x5, a sixth weight index x6, a seventh weight index x7, an
eighth
weight index x8 and a ninth weight index x9 are each evaluated as a decimal
between
0 and 1;
summate and average the respective weight indices of the number of sharings,
the
number of praisings, the ni mber of commentings, the proportion of positive
comments, the number of listings as favorites and the number of followings;
multiply a result of the summating and a result of the averaging with the
respective
weight indices of the released video playback integrity rates;
calculate to obtain the video quality scores of the at least one uploader;
sort category proportions of released videos of the at least one uploader
within the
second preset period in combination with the time decay;
map the sorted category proportions to a range of [x10,1] wherein x10 is a
tenth
weight;
determine respective weight indices of the category proportions, wherein a
tenth
weight index x10 is evaluated as a decimal between 0 and 1;
Date Recue/Date Received 2023-11-21

multiply the respective weight indices of the category proportions; and
calculate to obtain the video verticality scores of the at least one uploader.
45. The memory of claim 44 wherein obtaining the first weight index xl, the
second weight
index x2, the third weight index x3, the fourth weight index x4, the fifth
weight index x5,
the sixth weight index x6, the seventh weight index x7, the eighth weight
index x8, the
ninth weight index x9 and the tenth weight index x10 comprises:
taking, respectively, at least one dimension feature score to which the first
weight
index xl, the second weight index x2, the third weight index x3, the fourth
weight
index x4, the fifth weight index x5, the sixth weight index x6, the seventh
weight
index x7, the eighth weight index x8, the ninth weight index x9 and the tenth
weight
index x10 correspond as independent variables;
taking at least one following degree of the at least one uploader after
exposure as
dependent variables; and
employing a RandomForest algorithm and a GBDT algorithm to calculate the first

weight index xl, the second weight index x2, the third weight index x3, the
fourth
weight index x4, the fifth weight index x5, the sixth weight index x6, the
seventh
weight index x7, the eighth weight index x8, the ninth weight index x9 and the
tenth
weight index x10.
46. The memory of any one of claims 42 to 45 wherein the computer is
further configured to:
multiply the released video activity scores, the video quality scores and the
video
verticality scores; and
calculate to obtain the at least one comprehensive score value of the at least
one
uploader.
47. The memory of any one of claims 41 to 46 wherein the computer is
further configured to:
56
Date Recue/Date Received 2023-11-21

eliminate one or more of hotspot videos and miss-clicked videos from the user
played-back video data;
count target user tags whose number of videos occupies a proportion that is
not lower
than a preset proportion according to the preset tagging rule wherein a preset

proportion is greater than one;
calculate the target user tags; and
generate corresponding user word vectors of the target user tags.
48. The memory of any one of claims 41 to 47 wherein the computer is
further configured to
recommend the uploader information to the user.
49. The memory of any one of claims 41 to 48 wherein the computer is
further configured to
push a video of the at least one video category tag corresponding to the
uploader word
vector result to the user.
50. The memory of any one of claims 41 to 49 wherein one or more of the
activity of the at
least one uploader, quality of the at least one uploader, and verticality of
the at least one
uploader are comprehensively considered to facilitate subsequent comprehensive
scoring
of a quality of the at least one uploader.
51. A device for uploader matching comprising:
a calculating module configured to:
obtain a released video data of at least one uploader;
determine at least one comprehensive score value of the at least one
uploader from at least one dimension feature score according to the
released video data; and
57
Date Recue/Date Received 2023-11-21

screen out at least one target uploader according to the at least one
comprehensive score value of the at least one uploader;
wherein the calculating module further comprises:
a first calculating sub-module configured to calculate at least one
score of one or more of at least one dimension feature in released
video activity scores of the at least one uploader, video quality scores
of the at least one uploader, and video verticality scores of the at least
one uploader according to the released video data;
a second calculating sub-module configured to calculate the at least
one comprehensive score value of the at least one uploader according
to the one or more dimension feature score; and
a screening sub-module configured to select the at least one uploader
who rank above a threshold to serve as the at least one target uploader
according to a sequence of the at least one comprehensive score value
of the at least one uploader arranged in a decreasing order, wherein the
threshold is an integer greater than one;
an uploader word vector generating module configured to:
make statistics on the released video data of the at least one target uploader

according to a preset tagging rule; and
generate a corresponding uploader word vector of at least one video
category tag;
a user word vector generating module configured to:
obtain a user played-back video data within a first preset period;
58
Date Recue/Date Received 2023-11-21

make statistics on the user played-back video data according to the preset
tagging rule; and
generate a corresponding user word vector of the at least one video
category tag; and
a matching module configured to:
match the corresponding of the at least one video category tag of the at
least one uploader word vector and a user word vector;
obtain an uploader word vector result that has reached a target matching
degree with the user; and
determine corresponding uploader information according to the uploader
word vector result.
52. The device of claim 51 wherein the first calculating sub-module
configured to:
check an external order information of each target order according to a preset
label
generating rule; and
obtain a checking result of the each target order.
53. The device according to any one of claims 51 to 52 wherein the first
calculating submodule
is further configured to:
sort a number of released videos of the at least one uploader within a second
preset
period, and a volume of released videos played back within the second preset
period
respectively in combination with a time decay;
59
Date Recue/Date Received 2023-11-21

map the sorted number of released videos and the sorted volume of released
videos
played back within the second preset period to a range of [x1,1] and a range
of
[x2,1], wherein x I is a first weight index and x2 is a second weight index
each
evaluated as a decimal between 0 and 1;
determine respective weight indices of the number of released videos and the
volume
of released videos played back;
multiply the respective weight indices of the number of released videos with
the
respective weight indices of the volume of released videos played back;
calculate to obtain the released video activity scores of the at least one
uploader;
sort one or more of a number of sharings, a number of praisings, a number of
commentings, a proportion of positive comments, a number of listings as
favorites, a
number of followings and released video playback integrity rates of released
videos
of the at least one uploader within the second preset period respectively
combined
with the time decay;
map the sorted number of sharings, number of praisings, number of commentings,

proportion of positive comments, number of listings as favorites, number of
followings and released video playback integrity rates to a range of [x3,1], a
range of
[x4,1], a range of [x5,1], a range of [x6,1], a range of [x7,1], a range of
[x8,1] and a
range of [x9,1], respectively wherein a x3 is a third weight index, x4 is a
fourth
weight index, x5 is a fifth weight index, x6 is a sixth weight index, x7 is a
seventh
weight index, x8 is an eighth weight index and x9 a ninth weight index;
Date Recue/Date Received 2023-11-21

determine respective weight indices of the number of sharings, the number of
praisings, the number of commentings, the proportion of positive comments, the

number of listings as favorites, the number of followings and the released
video
playback integrity rates, wherein a third weight index x3, a fourth weight
index x4, a
fifth weight index x5, a sixth weight index x6, a seventh weight index x7, an
eighth
weight index x8 and a ninth weight index x9 are each evaluated as a decimal
between
0 and 1;
summate and average the respective weight indices of the number of sharings,
the
number of praisings, the number of commentings, the proportion of positive
comments, the number of listings as favorites and the number of followings;
multiply a result of the summating and a result of the averaging with the
respective
weight indices of the released video playback integrity rates;
calculate to obtain the video quality scores of the at least one uploader;
sort category proportions of released videos of the at least one uploader
within the
second preset period in combination with the time decay;
map the sorted category proportions to a range of [x10,1] wherein x10 is a
tenth
weight;
determine respective weight indices of the category proportions, wherein a
tenth
weight index x10 is evaluated as a decimal between 0 and 1;
multiply the respective weight indices of the category proportions; and
calculate to obtain the video verticality scores of the at least one uploader.
54.
The device of claim 53 wherein obtaining the first weight index xl, the second
weight
index x2, the third weight index x3, the fourth weight index x4, the fifth
weight index x5,
the sixth weight index x6, the seventh weight index x7, the eighth weight
index x8, the
ninth weight index x9 and the tenth weight index x10 comprises:
61
Date Reçue/Date Received 2023-11-21

taking, respectively, at least one dimension feature score to which the first
weight
index xl, the second weight index x2, the third weight index x3, the fourth
weight
index x4, the fifth weight index x5, the sixth weight index x6, the seventh
weight
index x7, the eighth weight index x8, the ninth weight index x9 and the tenth
weight
index x10 correspond as independent variables;
taking at least one following degree of the at least one uploader after
exposure as
dependent variables; and
employing a RandomForest algorithm and a GBDT algorithm to calculate the first

weight index xl, the second weight index x2, the third weight index x3, the
fourth
weight index x4, the fifth weight index x5, the sixth weight index x6, the
seventh
weight index x7, the eighth weight index x8, the ninth weight index x9 and the
tenth
weight index x10.
55. The device of any one of claims 51 to 54 wherein the second calculating
sub-module is
configured to:
multiply the released video activity scores, the video quality scores and the
video
verticality scores; and
calculate to obtain the at least one comprehensive score value of the at least
one
uploader.
56. The device of any one of claims 51 to 55 wherein the user word vector
generating module
is further configured to:
eliminate one or more of hotspot videos and miss-clicked videos from the user
played-back video data;
count target user tags whose number of videos occupies a proportion that is
not lower
than a preset proportion according to the preset tagging rule wherein a preset

proportion is greater than one;
62
Date Recue/Date Received 2023-11-21

calculate the target user tags; and
generate corresponding user word vectors of the target user tags.
57. The device of any one of claims 51 to 56 further comprising a data
recommending module
configured to recommend the uploader information to the user.
58. The device of any one of claims 51 to 57 further comprising the data
recommending module
configured to push a video of the at least one video category tag
corresponding to the
uploader word vector result to the user.
59. The device of any one of claims 51 to 58 wherein one or more of
activity of the at least one
uploader, quality of the at least one uploader, and verticality of the at
least one uploader
are comprehensively considered to facilitate subsequent comprehensive scoring
of a quality
of the at least one uploader.
60. A system for uploader matching comprising:
a calculating module configured to:
obtain a released video data of at least one uploader;
determine at least one comprehensive score value of the at least one
uploader from at least one dimension feature score according to the
released video data; and
screen out at least one target uploader according to the comprehensive
score values of the uploader;
wherein the calculating module further comprises:
63
Date Recue/Date Received 2023-11-21

a first calculating sub-module configured to calculate at least one
score of one or more of at least one dimension feature in released
video activity scores of the at least one uploader, video quality scores
of the at least one uploader, and video verticality scores of the at least
one uploader according to the released video data;
a second calculating sub-module configured to calculate the at least
one comprehensive score value of the at least one uploader according
to the one or more dimension feature score; and
a screening sub-module configured to select the at least one uploader
who rank above a threshold to serve as the at least one target uploader
according to a sequence of the at least one comprehensive score value
of the at least one uploaderthe at least one uploader arranged in a
decreasing order, wherein the threshold is an integer greater than one;
an uploader word vector generating module configured to:
make statistics on the released video data of the at least one target uploader

according to a preset tagging rule; and
generate a corresponding uploader word vector of at least one video
category tag;
a user word vector generating module configured to:
obtain a user played-back video data within a first preset period;
make statistics on the user played-back video data according to the preset
tagging rule; and
generate a corresponding user word vector of the at least one video
category tag; and
64
Date Recue/Date Received 2023-11-21

a matching module configured to:
match the corresponding of the at least one video category tag of the at
least one uploader word vector and a user word vector;
obtain an uploader word vector result that has reached a target matching
degree with the user; and
determine corresponding uploader infonnation according to the uploader
word vector result.
61. The system of claim 60 wherein the first calculating sub-module
configured to:
check an external order information of each target order according to a preset
label
generating nile; and
obtain a checking result of the each target order.
62. The system according to any one of claims 60 to 61 wherein the first
calculating submodule
is further configured to:
sort a number of released videos of the at least one uploader within a second
preset
period, and a volume of released videos played back within the second preset
period
respectively in combination with a time decay;
map the sorted number of released videos and the sorted volume of released
videos
played back within the second preset period to a range of [x1,1] and a range
of
[x2,1], wherein x 1 is a first weight index and x2 is a second weight index
each
evaluated as a decimal between 0 and 1;
determine respective weight indices of the number of released videos and the
volume
of released videos played back;
Date Recue/Date Received 2023-11-21

multiply the respective weight indices of the number of released videos with
the
respective weight indices of the volume of released videos played back;
calculate to obtain the released video activity scores of the at least one
uploader;
sort one or more of a number of sharings, a number of praisings, a number of
commentings, a proportion of positive comments, a number of listings as
favorites, a
number of followings and released video playback integrity rates of released
videos
of the at least one uploader within the second preset period respectively
combined
with the time decay;
map the sorted number of sharings, number of praisings, number of commentings,

proportion of positive comments, number of listings as favorites, number of
followings and released video playback integrity rates to a range of [x3,1], a
range of
[x4,1], a range of [x5,1], a range of [x6,1], a range of [x7,1], a range of
[x8,1] and a
range of [x9,1], respectively wherein a x3 is a third weight index, x4 is a
fourth
weight index, x5 is a fifth weight index, x6 is a sixth weight index, x7 is a
seventh
weight index, x8 is an eighth weight index and x9 a ninth weight index;
determine respective weight indices of the number of sharings, the number of
praisings, the number of commentings, the proportion of positive comments, the

number of listings as favorites, the number of followings and the released
video
playback integiity rates, wherein a third weight index x3, a fourth weight
index x4, a
fifth weight index x5, a sixth weight index x6, a seventh weight index x7, an
eighth
weight index x8 and a ninth weight index x9 are each evaluated as a decimal
between
0 and 1;
summate and average the respective weight indices of the number of sharings,
the
number of praisings, the number of commentings, the proportion of positive
comments, the number of listings as favorites and the number of followings;
66
Date Reçue/Date Received 2023-11-21

multiply a result of the summating and a result of the averaging with the
respective
weight indices of the released video playback integrity rates;
calculate to obtain the video quality scores of the at least one uploader;
sort category proportions of released videos of the at least one uploader
within the
second preset period in combination with the time decay;
map the sorted category proportions to a range of [x10,1] wherein x10 is a
tenth
weight;
determine respective weight indices of the category proportions, wherein a
tenth
weight index x10 is evaluated as a decimal between 0 and 1;
multiply the respective weight indices of the category proportions; and
calculate to obtain the video verticality scores of the at least one uploader.
63.
The system of claim 62 wherein obtaining the first weight index xl, the second
weight
index x2, the third weight index x3, the fourth weight index x4, the fifth
weight index x5,
the sixth weight index x6, the seventh weight index x7, the eighth weight
index x8, the
ninth weight index x9 and the tenth weight index x10 comprises:
taking, respectively, at least one dimension feature score to which the first
weight
index xl, the second weight index x2, the third weight index x3, the fourth
weight
index x4, the fifth weight index x5, the sixth weight index x6, the seventh
weight
index x7, the eighth weight index x8, the ninth weight index x9 and the tenth
weight
index x10 correspond as independent variables;
taking at least one following degree of the at least one uploader after
exposure as
dependent variables; and
67
Date Recue/Date Received 2023-11-21

employing a RandomForest algorithm and a GBDT algorithm to calculate the first

weight index xl, the second weight index x2, the third weight index x3, the
fourth
weight index x4, the fifth weight index x5, the sixth weight index x6, the
seventh
weight index x7, the eighth weight index x8, the ninth weight index x9 and the
tenth
weight index x10.
64. The system of any one of claims 60 to 63 wherein the second calculating
sub-module is
configured to:
multiply the released video activity scores, the video quality scores and the
video
verticality scores; and
calculate to obtain the at least one comprehensive score value of the at least
one
uploader.
65. The system of any one of claims 60 to 64 wherein the user word vector
generating module
is further configured to:
eliminate one or more of hotspot videos and miss-clicked videos from the user
played-back video data;
count target user tags whose number of videos occupies a proportion that is
not lower
than a preset proportion according to the preset tagging rule wherein a preset

proportion is greater than one;
calculate the target user tags; and
generate corresponding user word vectors of the target user tags.
66. The system of any one of claims 60 to 65 further comprising a data
recommending module
configured to recommend the uploader information to the user.
68
Date Recue/Date Received 2023-11-21

67. The system of any one of claims 60 to 66 further comprising the data
recommending
module configured to push a video of the at least one video category tag
corresponding to
the uploader word vector result to the user.
68. The system of any one of claims 60 to 67 wherein one or more of
activity of the at least
one uploader, quality of the at least one uploader, and verticality of the at
least one uploader
are comprehensively considered to facilitate subsequent comprehensive scoring
of a quality
of the at least one uploader.
69. A method for uploader matching comprising:
obtaining a released video data of at least one uploader;
determining at least one comprehensive score value of the at least one
uploader from
at least one dimension feature score according to the released video data;
calculating at least one score of one or more of at least one dimension
feature in
released video activity scores of the at least one uploader, video quality
scores of the
at least one uploader, and video verticality scores of the at least one
uploader
according to the released video data;
calculating the at least one comprehensive score value of the at least one
uploader
according to the one or more dimension feature score; and
selecting the at least one uploader who rank above a threshold to serve as the
at least
one target uploader according to a sequence of the at least one comprehensive
score
value of the at least one uploader arranged in a decreasing order, wherein the

threshold is an integer greater than one;
screening out at least one target uploader according to the at least one
comprehensive
score value of the at least one uploader;
making statistics on the released video data of the at least one target
uploader
according to a preset tagging rule; and
69
Date Recue/Date Received 2023-11-21

generating a corresponding uploader word vector of at least one video category
tag;
obtaining a user played-back video data within a first preset period;
making statistics on the user played-back video data according to the preset
tagging
rule; and
generating a corresponding user word vector of the at least one video category
tag;
and
matching the corresponding of the at least one video category tag of the at
least one
uploader word vector and a user word vector;
obtaining an uploader word vector result that has reached a target matching
degree
with the user; and
determining corresponding uploader information according to the uploader word
vector result.
70. The method of claim 69 further comprising:
checking an external order information of each target order according to a
preset label
generating rule; and
obtaining a checking result of the each target order.
71. The method according to any one of claims 69 to 70 further comprising:
sorting a number of released videos of the at least one uploader within a
second
preset period, and a volume of released videos played back within the second
preset
period respectively in combination with a time decay;
Date Recue/Date Received 2023-11-21

mapping the sorted number of released videos and the sorted volume of released

videos played back within the second preset period to a range of [x1,1] and a
range
of [x2,1], wherein xl is a first weight index and x2 is a second weight index
each
evaluated as a decimal between 0 and 1;
determining respective weight indices of the number of released videos and the

volume of released videos played back;
multiplying the respective weight indices of the number of released videos
with the
respective weight indices of the volume of released videos played back;
calculating to obtain the released video activity scores of the at least one
uploader;
sorting one or more of a number of sharings, a number of praisings, a number
of
commentings, a proportion of positive comments, a number of listings as
favorites, a
number of followings and released video playback integrity rates of released
videos
of the at least one uploader within the second preset period respectively
combined
with the time decay;
mapping the sorted number of sharings, number of praisings, number of
commentings, proportion of positive comments, number of listings as favorites,

number of followings and released video playback integrity rates to a range of
[x3,1],
a range of [x4,1], a range of [x5,1], a range of [x6,1], a range of [x7,1], a
range of
[x8,1] and a range of [x9,1], respectively wherein a x3 is a third weight
index, x4 is a
fourth weight index, x5 is a fifth weight index, x6 is a sixth weight index,
x7 is a
seventh weight index, x8 is an eighth weight index and x9 a ninth weight
index;
71
Date Recue/Date Received 2023-11-21

determining respective weight indices of the number of sharings, the number of

praisings, the number of commentings, the proportion of positive comments, the

number of listings as favorites, the number of followings and the released
video
playback integrity rates, wherein a third weight index x3, a fourth weight
index x4, a
fifth weight index x5, a sixth weight index x6, a seventh weight index x7, an
eighth
weight index x8 and a ninth weight index x9 are each evaluated as a decimal
between
0 and 1;
summating and averaging the respective weight indices of the number of
sharings,
the number of praisings, the number of commentings, the proportion of positive

comments, the number of listings as favorites and the number of followings;
multiplying a result of the summating and a result of the averaging with the
respective weight indices of the released video playback integrity rates;
calculating to obtain the video quality scores of the at least one uploader;
sorting category proportions of released videos of the at least one uploader
within the
second preset period in combination with the time decay;
mapping the sorted category proportions to a range of [x10,1] wherein x10 is a
tenth
weight;
determining respective weight indices of the category proportions, wherein a
tenth
weight index x10 is evaluated as a decimal between 0 and 1;
multiplying the respective weight indices of the category proportions; and
calculating to obtain the video verticality scores of the at least one
uploader.
72.
The method of claim 71 wherein obtaining the first weight index xl, the second
weight
index x2, the third weight index x3, the fourth weight index x4, the fifth
weight index x5,
the sixth weight index x6, the seventh weight index x7, the eighth weight
index x8, the
ninth weight index x9 and the tenth weight index x10 comprises:
72
Date Recue/Date Received 2023-11-21

taking, respectively, at least one dimension feature score to which the first
weight
index xl, the second weight index x2, the third weight index x3, the fourth
weight
index x4, the fifth weight index x5, the sixth weight index x6, the seventh
weight
index x7, the eighth weight index x8, the ninth weight index x9 and the tenth
weight
index x10 correspond as independent variables;
taking at least one following degree of the at least one uploader after
exposure as
dependent variables; and
employing a RandomForest algorithm and a GBDT algorithm to calculate the first

weight index xl, the second weight index x2, the third weight index x3, the
fourth
weight index x4, the fifth weight index x5, the sixth weight index x6, the
seventh
weight index x7, the eighth weight index x8, the ninth weight index x9 and the
tenth
weight index x10.
73. The method of any one of claims 69 to 72 further comprising:
multiplying the released video activity scores, the video quality scores and
the video
verticality scores; and
calculating to obtain the at least one comprehensive score value of the at
least one
uploader.
74. The method of any one of claims 69 to 73 further comprising:
eliminating one or more of hotspot videos and miss-clicked videos from the
user
played-back video data;
counting target user tags whose number of videos occupies a proportion that is
not
lower than a preset proportion according to the preset tagging rule wherein a
preset
proportion is greater than one;
calculating the target user tags; and
73
Date Recue/Date Received 2023-11-21

generating corresponding user word vectors of the target user tags.
75. The method of any one of claims 69 to 74 further comprising
recommending the uploader
information to the user.
76. The method of any one of claims 69 to 75 further comprising pushing a
video of the video
category tag corresponding to the uploader word vector result to the user.
77. The method of any one of claims 69 to 76 wherein one or more of the
activity of the at least
one uploader, quality of the at least one uploader, and verticality of the at
least one uploader
are comprehensively considered to facilitate subsequent comprehensive scoring
of a quality
of the at least one uploader.
78. A computer equipment for uploader matching comprising a computer
readable physical
memory and a processor communicatively connected to the memory wherein the
processor
is configured to execute a computer-executable instructions stored on the
memory and
wherein the processor when executing the computer-executable instructions is
configured
to:
obtain a released video data of at least one uploader;
determine at least one comprehensive score value of the at least one uploader
from at
least one dimension feature score according to the released video data;
calculate at least one score of one or more of at least one dimension feature
in
released video activity scores of the at least one uploader, video quality
scores of the
at least one uploader, and video verticality scores of the at least one
uploader
according to the released video data;
calculate the at least one comprehensive score value of the at least one
uploader
according to the one or more dimension feature score;
74
Date Recue/Date Received 2023-11-21

select the at least one uploader who rank above a threshold to serve as the at
least one
target uploader according to a sequence of the at least one comprehensive
score value
of the at least one uploader arranged in a decreasing order, wherein the
threshold is
an integer greater than one;
screen out at least one target uploader according to the comprehensive score
values
of the uploader;
make statistics on the released video data of the at least one target uploader
according
to a preset tagging rule; and
generate a corresponding uploader word vector of at least one video category
tag;
obtain a user played-back video data within a first preset period;
make statistics on the user played-back video data according to the preset
tagging
rule; and
generate a corresponding user word vector of the at least one video category
tag; and
match the corresponding of the at least one video category tag of the at least
one
uploader word vector and the user word vector;
obtain an uploader word vector result that has reached a target matching
degree with
the user; and
determine corresponding uploader information according to the uploader word
vector
result.
79. The computer equipment of claim 78 wherein the processor is further
configured to:
check an external order information of each target order according to a preset
label
generating rule; and
obtain a checking result of the each target order.
Date Recue/Date Received 2023-11-21

80.
The computer equipment according to any one of claims 78 to 79 wherein the
processor is
further configured to:
sort a number of released videos of the at least one uploader within a second
preset
period, and a volume of released videos played back within the second preset
period
respectively in combination with a time decay;
map the sorted number of released videos and the sorted volume of released
videos
played back within the second preset period to a range of [x1,1] and a range
of
[x2,1], wherein x 1 is a first weight index and x2 is a second weight index
each
evaluated as a decimal between 0 and 1;
determine respective weight indices of the number of released videos and the
volume
of released videos played back;
multiply the respective weight indices of the number of released videos with
the
respective weight indices of the volume of released videos played back;
calculate to obtain the released video activity scores of the at least one
uploader;
sort one or more of a number of sharings, a number of praisings, a number of
commentings, a proportion of positive comments, a number of listings as
favorites, a
number of followings and released video playback integrity rates of released
videos
of the at least one uploader within the second preset period respectively
combined
with the time decay;
map the sorted number of sharings, number of praisings, number of commentings,

proportion of positive comments, number of listings as favorites, number of
followings and released video playback integrity rates to a range of [x3,1], a
range of
[x4,1], a range of [x5,1], a range of [x6,1], a range of [x7,1], a range of
[x8,1] and a
range of [x9,1], respectively wherein a x3 is a third weight index, x4 is a
fourth
weight index, x5 is a fifth weight index, x6 is a sixth weight index, x7 is a
seventh
weight index, x8 is an eighth weight index and x9 a ninth weight index;
76
Date Recue/Date Received 2023-11-21

determine respective weight indices of the number of sharings, the number of
praisings, the number of commentings, the proportion of positive comments, the

number of listings as favorites, the number of followings and the released
video
playback integrity rates, wherein a third weight index x3, a fourth weight
index x4, a
fifth weight index x5, a sixth weight index x6, a seventh weight index x7, an
eighth
weight index x8 and a ninth weight index x9 are each evaluated as a decimal
between
0 and 1;
summate and average the respective weight indices of the number of sharings,
the
number of praisings, the number of commentings, the proportion of positive
comments, the number of listings as favorites and the number of followings;
multiply a result of the summating and a result of the averaging with the
respective
weight indices of the released video playback integrity rates;
calculate to obtain the video quality scores of the at least one uploader;
sort category proportions of released videos of the at least one uploader
within the
second preset period in combination with the time decay;
map the sorted category proportions to a range of [x10,1] wherein x10 is a
tenth
weight;
determine respective weight indices of the category proportions, wherein a
tenth
weight index x10 is evaluated as a decimal between 0 and 1;
multiply the respective weight indices of the category proportions; and
calculate to obtain the video verticality scores of the at least one uploader.
81.
The computer equipment of claim 80 wherein obtaining the first weight index
xl, the
second weight index x2, the third weight index x3, the fourth weight index x4,
the fifth
weight index x5, the sixth weight index x6, the seventh weight index x7, the
eighth weight
index x8, the ninth weight index x9 and the tenth weight index x10 comprises:
77
Date Recue/Date Received 2023-11-21

taking, respectively, at least one dimension feature score to which the first
weight
index xl, the second weight index x2, the third weight index x3, the fourth
weight
index x4, the fifth weight index x5, the sixth weight index x6, the seventh
weight
index x7, the eighth weight index x8, the ninth weight index x9 and the tenth
weight
index x10 correspond as independent variables;
taking at least one following degree of the at least one uploader after
exposure as
dependent variables; and
employing a RandomForest algorithm and a GBDT algorithm to calculate the first

weight index xl, the second weight index x2, the third weight index x3, the
fourth
weight index x4, the fifth weight index x5, the sixth weight index x6, the
seventh
weight index x7, the eighth weight index x8, the ninth weight index x9 and the
tenth
weight index x10.
82. The computer equipment of any one of claims 78 to 81 wherein the
processor is further
configured to:
multiply the released video activity scores, the video quality scores and the
video
verticality scores; and
calculate to obtain the at least one comprehensive score value of the at least
one
uploader.
83. The computer equipment of any one of claims 78 to 82 wherein the
processor is further
configured to:
eliminate one or more of hotspot videos and miss-clicked videos from the user
played-back video data;
count target user tags whose number of videos occupies a proportion that is
not lower
than a preset proportion according to the preset tagging rule wherein a preset

proportion is greater than one;
78
Date Recue/Date Received 2023-11-21

calculate the target user tags; and
generate corresponding user word vectors of the target user tags.
84. The computer equipment of any one of claims 78 to 83 wherein the
processor is further
configured to recommend the uploader information to the user.
85. The computer equipment of any one of claims 78 to 84 wherein the
processor is further
configured to push a video of the at least one video category tag
corresponding to the
uploader word vector result to the user.
86. The computer equipment of any one of claims 78 to 85 wherein one or
more of the activity
of the at least one uploader, quality of the at least one uploader, and
verticality of the at
least one uploader are comprehensively considered to facilitate subsequent
comprehensive
scoring of a quality of the at least one uploader.
87. A computer readable physical memory having stored upon it a computer-
executable
instructions when executed by a computer configured to:
obtain a released video data of at least one uploader;
determine at least one comprehensive score value of the at least one uploader
from at
least one dimension feature score according to the released video data;
calculate at least one score of one or more of at least one dimension feature
in
released video activity scores of the at least one uploader, video quality
scores of the
at least one uploader, and video verticality scores of the at least one
uploader
according to the released video data;
calculate the at least one comprehensive score value of the at least one
uploader
according to the one or more dimension feature score; and
79
Date Recue/Date Received 2023-11-21

select the at least one uploader who rank above a threshold to serve as the at
least one
target uploader according to a sequence of the at least one comprehensive
score value
of the at least one uploader arranged in a decreasing order, wherein the
threshold is
an integer greater than one;
screen out at least one target uploader according to the comprehensive score
values
of the uploader;
make statistics on the released video data of the at least one target uploader
according
to a preset tagging rule; and
generate a corresponding uploader word vector of at least one video category
tag;
obtain a user played-back video data within a first preset period;
make statistics on the user played-back video data according to the preset
tagging
rule; and
generate a corresponding user word vector of the at least one video category
tag; and
match the corresponding of the at least one video category tag of the at least
one
uploader word vector and a user word vector;
obtain an uploader word vector result that has reached a target matching
degree with
the user; and
determine corresponding uploader information according to the uploader word
vector
result.
88. The memory of claim 87 wherein the computer is further configured to:
check an external order information of each target order according to a preset
label
generating rule; and
obtain a checking result of the each target order.
Date Recue/Date Received 2023-11-21

89. The memory any one of claims 87 to 88 wherein the processor is further
configured to:
sort a number of released videos of the at least one uploader within a second
preset
period, and a volume of released videos played back within the second preset
period
respectively in combination with a time decay;
map the sorted number of released videos and the sorted volume of released
videos
played back within the second preset period to a range of [x1,1] and a range
of
[x2,1], wherein x I is a first weight index and x2 is a second weight index
each
evaluated as a decimal between 0 and 1;
determine respective weight indices of the number of released videos and the
volume
of released videos played back;
multiply the respective weight indices of the number of released videos with
the
respective weight indices of the volume of released videos played back;
calculate to obtain the released video activity scores of the at least one
uploader;
sort one or more of a number of sharings, a number of praisings, a number of
commentings, a proportion of positive comments, a number of listings as
favorites, a
number of followings and released video playback integrity rates of released
videos
of the at least one uploader within the second preset period respectively
combined
with the time decay;
map the sorted number of sharings, number of praisings, number of commentings,

proportion of positive comments, number of listings as favorites, number of
followings and released video playback integrity rates to a range of [x3,1], a
range of
[x4,1], a range of [x5,1], a range of [x6,1], a range of [x7,1], a range of
[x8,1] and a
range of [x9,1], respectively wherein a x3 is a third weight index, x4 is a
fourth
weight index, x5 is a fifth weight index, x6 is a sixth weight index, x7 is a
seventh
weight index, x8 is an eighth weight index and x9 a ninth weight index;
81
Date Recue/Date Received 2023-11-21

determine respective weight indices of the number of sharings, the number of
praisings, the number of commentings, the proportion of positive comments, the

number of listings as favorites, the number of followings and the released
video
playback integrity rates, wherein a third weight index x3, a fourth weight
index x4, a
fifth weight index x5, a sixth weight index x6, a seventh weight index x7, an
eighth
weight index x8 and a ninth weight index x9 are each evaluated as a decimal
between
0 and 1;
summate and average the respective weight indices of the number of sharings,
the
number of praisings, the number of commentings, the proportion of positive
comments, the number of listings as favorites and the number of followings;
multiply a result of the summating and a result of the averaging with the
respective
weight indices of the released video playback integrity rates;
calculate to obtain the video quality scores of the at least one uploader;
sort category proportions of released videos of the at least one uploader
within the
second preset period in combination with the time decay;
map the sorted category proportions to a range of [x10,1] wherein x10 is a
tenth
weight;
determine respective weight indices of the category proportions, wherein a
tenth
weight index x10 is evaluated as a decimal between 0 and 1;
multiply the respective weight indices of the category proportions; and
calculate to obtain the video verticality scores of the at least one uploader.
90.
The memory of claim 89 wherein obtaining the first weight index xl, the second
weight
index x2, the third weight index x3, the fourth weight index x4, the fifth
weight index x5,
the sixth weight index x6, the seventh weight index x7, the eighth weight
index x8, the
ninth weight index x9 and the tenth weight index x10 comprises:
82
Date Recue/Date Received 2023-11-21

taking, respectively, at least one dimension feature score to which the first
weight
index xl, the second weight index x2, the third weight index x3, the fourth
weight
index x4, the fifth weight index x5, the sixth weight index x6, the seventh
weight
index x7, the eighth weight index x8, the ninth weight index x9 and the tenth
weight
index x10 correspond as independent variables;
taking at least one following degree of the at least one uploader after
exposure as
dependent variables; and
employing a RandomForest algorithm and a GBDT algorithm to calculate the first

weight index xl, the second weight index x2, the third weight index x3, the
fourth
weight index x4, the fifth weight index x5, the sixth weight index x6, the
seventh
weight index x7, the eighth weight index x8, the ninth weight index x9 and the
tenth
weight index x10.
91. The memory of any one of claims 87 to 90 wherein the computer is
further configured to:
multiply the released video activity scores, the video quality scores and the
video
verticality scores; and
calculate to obtain the at least one comprehensive score value of the at least
one
uploader.
92. The memory of any one of claims 87 to 91 wherein the computer is
further configured to:
eliminate one or more of hotspot videos and miss-clicked videos from the user
played-back video data;
count target user tags whose number of videos occupies a proportion that is
not lower
than a preset proportion according to the preset tagging rule wherein a preset

proportion is greater than one;
calculate the target user tags; and
83
Date Recue/Date Received 2023-11-21

generate corresponding user word vectors of the target user tags.
93. The memory of any one of claims 87 to 92 wherein the computer is
further configured to
recommend the uploader information to the user.
94. The memory of any one of claims 87 to 93 wherein the computer is
further configured to
push a video of the at least one video category tag corresponding to the
uploader word
vector result to the user.
95. The memory of any one of claims 87 to 94 wherein one or more of the
activity of the at
least one uploader, quality of the at least one uploader, and verticality of
the at least one
uploader are comprehensively considered to facilitate subsequent comprehensive
scoring
of a quality of the at least one uploader.
84
Date Recue/Date Received 2023-11-21

Description

UPLOADER MATCHING METHOD AND DEVICE
BACKGROUND OF THE INVENTION
Technical Field
[0001] The present invention relates to the field of computer information
technology, and more
particularly to an uploader matching method and an uploader matching device.
Description of Related Art
[0002] Faced with massive amounts of video resources and users numbering in
the hundred
millions in the field of short video recommendation, it is of crucial
importance as how to
recommend high-quality videos preferred by users to target users, so as to
solve the
problem of information overload and to enhance time duration of stay and
satisfaction of
users. Qualified uploaders (persons who upload video and audio files onto
video websites,
forums, and ftp sites) are high-quality video releasers proven by many users,
the
recommendation of qualified uploaders to users with similar user portraits
enables the
users to more conveniently acquire quality videos of interest to them, and
such
recommendation will increase user adhesion and satisfaction to a greater
extent. How to
qualitatively and quantitatively evaluate the qualities of short video
uploaders directly
decides whether most similar and high-quality uploaders can be recommended
precisely
to target users.
[0003] Currently, the number of users registered in some video platforms
reaches hundred
millions, with daily UV accesses exceeding ten millions, and the volume of
plays per day
is even higher in mobile ends. In order that users find out contents of
interest to them
from massive videos, precise user portraits and the recommending system exert
significant functions. Once qualified uploaders are recommended to and
followed by
most similar target users, it will be made possible for the users to
consistently watch high-
quality videos of possible interest to them.
1
CA 03150500 2022-3-8

SUMMARY OF THE INVENTION
[0004] In order to overcome problems pending in the state of the art,
embodiments of the present
invention provide an uploader matching method and an uploader matching device,

whereby comprehensive evaluation of uploaders is achieved through a
comprehensive
scoring scheme with multiple data sources and multi-dimensional variables,
high-quality
uploader information is extracted therefrom, and matched with user word vector
that is
realized through user portrait, and high-quality uploaders are finally
recommended to
target users with high matching degrees, so that CTR (Click-Through-Rate),
user
playback volume and average playback integrity are enhanced, and user
experience is
enhanced.
[0005] The technical solutions are as follows.
[0006] According to one aspect, there is provided an uploader matching method
that comprises:
[0007] obtaining released video data of uploaders, determining comprehensive
score values of
the uploaders from one or more dimension feature score(s) according to the
released video
data, and screening out target uploaders according to the comprehensive score
values of
the uploaders;
[0008] making statistics on the released video data of the target uploaders
according to a preset
tagging rule, and generating corresponding uploader word vector of one or more
video
category tag(s);
[0009] obtaining user played-back video data within a first preset period,
making statistics on
the user played-back video data according to the preset tagging rule, and
generating
corresponding user word vector of the one or more video category tag(s); and
[0010] correspondingly matching the corresponding video category tags of the
uploader word
vector and the user word vector, obtaining an uploader word vector result that
has reached
a target matching degree with the user, and determining corresponding uploader
2
CA 03150500 2022-3-8

information according to the uploader word vector result.
[0011] Further, the step of determining comprehensive score values of the
uploaders from one
or more dimension feature score(s) according to the released video data, and
screening
out target uploaders according to the comprehensive score values of the
uploaders
includes:
[0012] calculating score(s) of one or more dimension feature(s) in released
video activity scores
of the uploaders, video quality scores of the uploaders, and video verticality
scores of the
uploaders according to the released video data;
[0013] calculating the comprehensive score values of the uploaders according
to the one or more
dimension feature score(s); and
[0014] selecting the uploaders who rank top N to serve as the target uploaders
according to a
sequence of the comprehensive score values of the uploaders arranged in a
decreasing
order, wherein N is an integer greater than 1.
[0015] Further, the steps of calculating score(s) of one or more dimension
feature(s) in released
video activity scores of the uploaders, video quality scores of the uploaders,
and video
verticality scores of the uploaders according to the released video data;
calculating the
comprehensive score values of the uploaders according to the one or more
dimension
feature score(s) include:
[0016] sorting the number of released videos of the uploaders within a second
preset period, and
a volume of released videos played back within the second preset period
respectively in
combination with time decay, mapping the sorted number of released videos and
the
sorted volume of released videos played back within the second preset period
to a range
of [x1,1] and a range of [x2,1], determining respective weight indices of the
number of
released videos and the volume of released videos played back, wherein a first
weight
index xl and a second weight index x2 are each evaluated as a decimal between
0 and 1,
thereafter multiplying the respective weight indices of the number of released
videos with
the respective weight indices of the volume of released videos played back,
and
3
CA 03150500 2022-3-8

calculating to obtain the released video activity scores of the uploaders;
[0017] sorting the number of sharings, the number of praisings, the number of
commentings, a
proportion of positive comments, the number of listings as favorites, the
number of
followings and released video playback integrity rates of released videos of
the uploaders
within the second preset period respectively combined with time decay, mapping
the
sorted number of sharings, number of praisings, number of commentings,
proportion of
positive comments, number of listings as favorites, number of followings and
released
video playback integrity rates to a range of [x3,1], a range of [x4,1], a
range of [x5,1], a
range of [x6,1], a range of [x7,1], a range of [x8,1] and a range of [x9,1],
respectively
determining respective weight indices of the number of sharings, the number of
praisings,
the number of commentings, the proportion of positive comments, the number of
listings
as favorites, the number of followings and the released video playback
integrity rates,
wherein a third weight index x3, a fourth weight index x4, a fifth weight
index x5, a sixth
weight index x6, a seventh weight index x7, an eighth weight index x8 and a
ninth weight
index x9 are each evaluated as a decimal between 0 and 1, thereafter summating
and
averaging the respective weight indices of the number of sharings, the number
of
praisings, the number of commentings, the proportion of positive comments, the
number
of listings as favorites and the number of followings, subsequently
multiplying the
summating and averaging result with the respective weight indices of the
released video
playback integrity rates, and calculating to obtain the video quality scores
of the uploaders;
[0018] sorting category proportions of released videos of the uploaders within
the second preset
period in combination with time decay, mapping the sorted category proportions
to a
range of [x10,1], determining respective weight indices of the category
proportions,
wherein a tenth weight index x10 is evaluated as a decimal between 0 and 1,
thereafter
multiplying the respective weight indices of the category proportions, and
calculating to
obtain the video verticality scores of the uploaders; and
[0019] multiplying the released video activity scores, the video quality
scores and the video
verticality scores, and calculating to obtain the comprehensive score values
of the
uploaders.
4
CA 03150500 2022-3-8

[0020] Moreover, the method further comprises: a step of obtaining the first
weight index xl, the
second weight index x2, the third weight index x3, the fourth weight index x4,
the fifth
weight index x5, the sixth weight index x6, the seventh weight index x7, the
eighth weight
index x8, the ninth weight index x9 and the tenth weight index x10:
[0021] respectively taking various dimension feature scores to which the first
weight index xl,
the second weight index x2, the third weight index x3, the fourth weight index
x4, the
fifth weight index x5, the sixth weight index x6, the seventh weight index x7,
the eighth
weight index x8, the ninth weight index x9 and the tenth weight index x10
correspond as
independent variables, taking following degrees of the uploaders after
exposure as
dependent variables, and employing a RandomForest algorithm and a GBDT
algorithm
to calculate the first weight index xl, the second weight index x2, the third
weight index
x3, the fourth weight index x4, the fifth weight index x5, the sixth weight
index x6, the
seventh weight index x7, the eighth weight index x8, the ninth weight index x9
and the
tenth weight index x10.
[0022] Further, the step of obtaining user played-back video data within a
first preset period,
making statistics on the user played-back video data according to the preset
tagging rule,
and generating corresponding user word vector of the one or more video
category tag(s)
includes:
[0023] eliminating hotspot videos and miss-clicked videos from the user played-
back video data,
counting out top N target user tags each of whose number of videos occupies a
proportion
that is not lower than a preset proportion according to the preset tagging
rule, calculating
the target user tags, and generating corresponding user word vectors of the
target user
tags, wherein N is an integer greater than 1.
[0024] Moreover, the method further comprises:
[0025] recommending the uploader information to the user; and/or
[0026] pushing a video of the video category tag corresponding to the uploader
word vector
CA 03150500 2022-3-8

result to the user.
[0027] According to another aspect, there is provided an uploader matching
device that
comprises:
[0028] a calculating module, for obtaining released video data of uploaders,
determining
comprehensive score values of the uploaders from one or more dimension feature
score(s)
according to the released video data, and screening out target uploaders
according to the
comprehensive score values of the uploaders;
[0029] an uploader word vector generating module, for making statistics on the
released video
data of the target uploaders according to a preset tagging rule, and
generating
corresponding uploader word vector of one or more video category tag(s);
[0030] a user word vector generating module, for obtaining user played-back
video data within
a first preset period, making statistics on the user played-back video data
according to the
preset tagging rule, and generating corresponding user word vector of the one
or more
video category tag(s); and
[0031] a matching module, for correspondingly matching the corresponding video
category tags
of the uploader word vector and the user word vector, obtaining an uploader
word vector
result that has reached a target matching degree with the user, and
determining
corresponding uploader information according to the uploader word vector
result.
[0032] Further, the calculating module includes a first calculating sub-
module, a second
calculating sub-module and a screening sub-module, of which:
[0033] the first calculating sub-module is employed for calculating score(s)
of one or more
dimension feature(s) in released video activity scores of the uploaders, video
quality
scores of the uploaders, and video verticality scores of' the uploaders
according to the
released video data;
[0034] the second calculating sub-module is employed for calculating the
comprehensive score
values of the uploaders according to the one or more dimension feature
score(s); and
[0035] the screening sub-module is employed for selecting the uploaders who
rank top N to serve
6
CA 03150500 2022-3-8

as the target uploaders according to a sequence of the comprehensive score
values of the
uploaders arranged in a decreasing order, wherein N is an integer greater than
1.
[0036] Further, the first calculating sub-module is employed for:
[0037] sorting the number of released videos of the uploaders within a second
preset period, and
a volume of released videos played back within the second preset period
respectively in
combination with time decay, mapping the sorted number of released videos and
the
sorted volume of released videos played back within the second preset period
to a range
of [x1,1] and a range of [x2,1], determining respective weight indices of the
number of
released videos and the volume of released videos played back, wherein a first
weight
index xl and a second weight index x2 are each evaluated as a decimal between
0 and 1,
thereafter multiplying the respective weight indices of the number of released
videos with
the respective weight indices of the volume of released videos played back,
and
calculating to obtain the released video activity scores of the uploaders;
[0038] sorting the number of sharings, the number of praisings, the number of
commentings, a
proportion of positive comments, the number of listings as favorites, the
number of
followings and released video playback integrity rates of released videos of
the uploaders
within the second preset period respectively combined with time decay, mapping
the
sorted number of sharings, number of praisings, number of commentings,
proportion of
positive comments, number of listings as favorites, number of followings and
released
video playback integrity rates to a range of [x3,1], a range of [x4,1], a
range of [x5,1], a
range of [x6,1], a range of [x7,1], a range of [x8,1] and a range of [x9,1],
respectively
determining respective weight indices of the number of sharings, the number of
praisings,
the number of commentings, the proportion of positive comments, the number of
listings
as favorites, the number of followings and the released video playback
integrity rates,
wherein a third weight index x3, a fourth weight index x4, a fifth weight
index x5, a sixth
weight index x6, a seventh weight index x7, an eighth weight index x8 and a
ninth weight
index x9 are each evaluated as a decimal between 0 and 1, thereafter summating
and
averaging the respective weight indices of the number of sharings, the number
of
7
CA 03150500 2022-3-8

praisings, the number of commentings, the proportion of positive comments, the
number
of listings as favorites and the number of followings, subsequently
multiplying the
summating and averaging result with the respective weight indices of the
released video
playback integrity rates, and calculating to obtain the video quality scores
of the uploaders;
and
[0039] sorting category proportions of released videos of the uploaders within
the second preset
period in combination with time decay, mapping the sorted category proportions
to a
range of [x10,1], determining respective weight indices of the category
proportions,
wherein a tenth weight index x10 is evaluated as a decimal between 0 and 1,
thereafter
multiplying the respective weight indices of the category proportions, and
calculating to
obtain the video verticality scores of the uploaders; and
[0040] the second calculating sub-module is employed for:
[0041] multiplying the released video activity scores, the video quality
scores and the video
verticality scores, and calculating to obtain the comprehensive score values
of the
uploaders.
[0042] Further, a step of obtaining the first weight index xl, the second
weight index x2, the
third weight index x3, the fourth weight index x4, the fifth weight index x5,
the sixth
weight index x6, the seventh weight index x7, the eighth weight index x8, the
ninth
weight index x9 and the tenth weight index x10 includes:
[0043] respectively taking various dimension feature scores to which the first
weight index xl,
the second weight index x2, the third weight index x3, the fourth weight index
x4, the
fifth weight index x5, the sixth weight index x6, the seventh weight index x7,
the eighth
weight index x8, the ninth weight index x9 and the tenth weight index x10
correspond as
independent variables, taking following degrees of the uploaders after
exposure as
dependent variables, and employing a RandomForest algorithm and a GBDT
algorithm
to calculate the first weight index xl, the second weight index x2, the third
weight index
x3, the fourth weight index x4, the fifth weight index x5, the sixth weight
index x6, the
seventh weight index x7, the eighth weight index x8, the ninth weight index x9
and the
8
CA 03150500 2022-3-8

tenth weight index x10.
[0044] Further, obtaining user played-back video data within a first preset
period, making
statistics on the user played-back video data according to the preset tagging
rule, and
generating corresponding user word vector of the one or more video category
tag(s)
includes:
[0045] eliminating hotspot videos and miss-clicked videos from the user played-
back video data,
counting out top N target user tags each of whose number of videos occupies a
proportion
that is not lower than a preset proportion according to the preset tagging
rule, calculating
the target user tags, and generating corresponding user word vectors of the
target user
tags, wherein N is an integer greater than 1.
[0046] Further, the device further comprises a data recommending module for
recommending
the uploader information to the user; and/or pushing a video of the video
category tag
corresponding to the uploader word vector result to the user.
[0047] The technical solutions provided by the embodiments of the present
invention bring about
the following advantageous effects.
[0048] 1. By sorting out in overall such data as released video information of
uploaders and user
behaviors, history records of user playbacks and such multi-dimensional
information as
relevant to the listing as favorites, sharing, praising, commenting and
playback integrities
of videos released by uploaders are obtained, a comprehensive scoring scheme
based on
multiple data sources and multi-dimensional variables achieves comprehensive
evaluation of uploaders, and high-quality uploader information is extracted
therefrom.
[0049] 2. Various dimension weights for evaluating uploader qualities are
obtained through
model training by corresponding algorithms, and evaluation of uploader
qualities is made
more precise.
9
CA 03150500 2022-3-8

[0050] 3. Uploaders are finely classified according to released video category
tags, whereby
precision of uploader portraits is enhanced.
[0051] 4. During the process of drawing portraits of users, hotspot videos and
possibly miss-
clicked videos are eliminated, portraits are drawn for the users by
differentiating different
interest tags, result sets are proportionally formed, and precision in
recommending result
sets is enhanced.
[0052] 5. Time decay is taken into consideration in the processes of
calculating user vector and
calculating uploader vector, whereby transfer of interest points is enhanced.
[0053] 6. By matching the high-quality uploader word vector with the user word
vector of a
precise user portrait, high-quality uploaders are finally recommended to
target users with
high matching degree, by following uploaders similar to themselves, users are
facilitated
to timely watch high-quality videos, whereby satisfaction of the users is
enhanced,
comparison report indicators are obtained through AB test, whereby CTR, user
playback
volume and average playback integrity are enhanced, and user experience is
enhanced as
a whole.
BRIEF DESCRIPTION OF THE DRAWINGS
[0054] To more clearly describe the technical solutions in the embodiments of
the present
invention, drawings required to illustrate the embodiments are briefly
introduced below.
Apparently, the drawings introduced below are merely directed to some
embodiments of
the present invention, while persons ordinarily skilled in the art may further
acquire other
drawings on the basis of these drawings without spending creative effort in
the process.
[0055] Fig. 1 is a flowchart illustrating the uploader matching method
provided by the
CA 03150500 2022-3-8

embodiments of the present invention;
[0056] Fig. 2 is a flowchart illustrating sub-steps of step 101 in Fig. 1;
[0057] Fig. 3 illustrates a preferred mode of execution for the setup of
dimension features of the
uploader comprehensive score values and score calculation thereof;
[0058] Fig. 4 illustrates a preferred mode of execution for the generation of
the uploader word
vector and the user word vector, and the corresponding tags matching; and
[0059] Fig. 5 is a view schematically illustrating the structure of the
uploader matching device
provided by the embodiments of the present invention.
DETAILED DESCRIPTION OF THE INVENTION
[0060] To make more lucid and clear the objectives, technical solutions and
advantages of the
present invention, the technical solutions in the embodiments of the present
invention will
be clearly and comprehensively described below with reference to the
accompanying
drawings in the embodiments of the present invention. Apparently, the
embodiments as
described are merely partial, rather than the entire, embodiments of the
present invention.
Any other embodiments makeable by persons ordinarily skilled in the art on the
basis of
the embodiments in the present invention without creative effort shall all
fall within the
protection scope of the present invention.
[0061] As should be noted, the terms "first" and "second" etc. are merely used
for the purpose
of description, and should not be understood to indicate or implicate relative
importance
or imply the number of technical features expressed thereby. Therefore, a
feature defined
by "first" or "second" can explicitly or implicitly include one or more such
feature(s).
The wording "a plurality of/plural" as used in the description of the present
invention
11
CA 03150500 2022-3-8

means "two or more" unless specifically defined otherwise.
[0062] In the uploader matching method and device provided by the embodiments
of the present
invention, by sorting out in overall such data as released video information
of uploaders
and user behaviors, history records of user playbacks and such multi-
dimensional
information as relevant to the listing as favorites, sharing, praising,
commenting and
playback integrities of videos released by uploaders are obtained, a
comprehensive
scoring scheme based on multiple data sources and multi-dimensional variables
achieves
comprehensive evaluation of uploaders, and high-quality uploader information
is
extracted therefrom; moreover, matching is performed with user word vector
realized
through user portrait, and high-quality uploaders are finally recommended to
target users
with high matching degrees, so that CTR (Click-Through-Rate), user playback
volume
and average playback integrity are enhanced, and user experience is enhanced.
Accordingly, the uploader matching method and device are applicable to such
application
scenarios as short video data processing, data matching or data pushing in the
field of
short video platforms.
[0063] The uploader matching method and device provided by the embodiments of
the present
invention are described in greater detail below in conjunction with specific
embodiments
and accompanying drawings.
[0064] Fig. 1 is a flowchart illustrating the uploader matching method
provided by the
embodiments of the present invention. Fig. 2 is a flowchart illustrating sub-
steps of step
101 in Fig. 1. Fig. 3 is a flowchart illustrating steps secondary to the sub-
step 1012 in Fig.
2.
[0065] As shown in Fig. 1, the uploader matching method provided by the
embodiments of the
present invention mainly comprises steps 101, 102, 103 and 104.
12
CA 03150500 2022-3-8

[0066] 101 - obtaining released video data of uploaders, determining
comprehensive score values
of the uploaders from one or more dimension feature score(s) according to the
released
video data, and screening out target uploaders according to the comprehensive
score
values of the uploaders.
[0067] Specifically, as shown in Fig. 2, step 101 may include the following
sub-steps:
[0068] 1011 ¨ obtaining released video data of uploaders;
[0069] 1012 ¨ calculating score(s) of one or more dimension feature(s) in
released video activity
scores of the uploaders, video quality scores of the uploaders, and video
verticality scores
of the uploaders according to the released video data;
[0070] 1013 ¨ calculating the comprehensive score values of the uploaders
according to the one
or more dimension feature score(s).
[0071] Specifically, the released video activity scores, the video quality
scores, and the video
verticality scores are multiplied to calculate and obtain the comprehensive
score values
of the uploaders.
[0072] 1014 - selecting the uploaders who rank top N to serve as the target
uploaders according
to a sequence of the comprehensive score values of the uploaders arranged in a
decreasing
order, wherein N is an integer greater than 1.
[0073] Specifically, sub-step 1012 further includes the following secondary
steps:
[0074] 1012a ¨ sorting the number of released videos of the uploaders within a
second preset
period, and a volume of released videos played back within the second preset
period
respectively in combination with time decay, mapping the sorted number of
released
videos and the sorted volume of released videos played back within the second
preset
period to a range of [x1,1] and a range of [x2,1], determining respective
weight indices
of the number of released videos and the volume of released videos played
back, wherein
a first weight index xi and a second weight index x2 are each evaluated as a
decimal
between 0 and 1, thereafter multiplying the respective weight indices of the
number of
released videos with the respective weight indices of the volume of released
videos played
13
CA 03150500 2022-3-8

back, and calculating to obtain the released video activity scores of the
uploaders;
[0075] 1012b ¨ sorting the number of sharings, the number of praisings, the
number of
commentings, a proportion of positive comments, the number of listings as
favorites, the
number of followings and released video playback integrity rates of released
videos of
the uploaders within the second preset period respectively combined with time
decay,
mapping the sorted number of sharings, number of praisings, number of
commentings,
proportion of positive comments, number of listings as favorites, number of
followings
and released video playback integrity rates to a range of [x3,1], a range of
[x4,1], a range
of [x5,1], a range of [x6,1], a range of [x7,1], a range of [x8,1] and a range
of [x9,1],
respectively determining respective weight indices of the number of sharings,
the number
of praisings, the number of commentings, the proportion of positive comments,
the
number of listings as favorites, the number of followings and the released
video playback
integrity rates, wherein a third weight index x3, a fourth weight index x4, a
fifth weight
index x5, a sixth weight index x6, a seventh weight index x7, an eighth weight
index x8
and a ninth weight index x9 are each evaluated as a decimal between 0 and 1,
thereafter
summating and averaging the respective weight indices of the number of
sharings, the
number of praisings, the number of commentings, the proportion of positive
comments,
the number of listings as favorites and the number of followings, subsequently

multiplying the summating and averaging result with the respective weight
indices of the
released video playback integrity rates, and calculating to obtain the video
quality scores
of the uploaders; and
[0076] 1012c - sorting category proportions of released videos of the
uploaders within the second
preset period in combination with time decay, mapping the sorted category
proportions
to a range of [x10,1], determining respective weight indices of the category
proportions,
wherein a tenth weight index x10 is evaluated as a decimal between 0 and 1,
thereafter
multiplying the respective weight indices of the category proportions, and
calculating to
obtain the video verticality scores of the uploaders.
[0077] The step of obtaining the first weight index xl, the second weight
index x2, the third
14
CA 03150500 2022-3-8

weight index x3, the fourth weight index x4, the fifth weight index x5, the
sixth weight
index x6, the seventh weight index x7, the eighth weight index x8, the ninth
weight index
x9 and the tenth weight index x10 is as follows:
[0078] respectively taking various dimension feature scores to which the first
weight index xl,
the second weight index x2, the third weight index x3, the fourth weight index
x4, the
fifth weight index x5, the sixth weight index x6, the seventh weight index x7,
the eighth
weight index x8, the ninth weight index x9 and the tenth weight index x10
correspond as
independent variables, taking following degrees of the uploaders after
exposure as
dependent variables, and employing a RandomForest algorithm and a GBDT
algorithm
to calculate the first weight index xl, the second weight index x2, the third
weight index
x3, the fourth weight index x4, the fifth weight index x5, the sixth weight
index x6, the
seventh weight index x7, the eighth weight index x8, the ninth weight index x9
and the
tenth weight index x10. The exposed following degrees of the uploaders are
aimed to
recommend the uploaders to target users, whether the users follow after
exposure is
specifically to take being followed as a positive sample of 1, and to take not
being
followed as a negative sample of 0, considering the problem of balance between
positive
and negative samples, the proportion between positive and negative samples is
usually
selected as 1:1 or 1:2.
[0079] It is possible to employ the RandomForest algorithm and the GBDT
algorithm to
construct a weight calculation model prior to the step of obtaining preset
weights of the
various dimension features, of course, in the case of not departing from the
inventive
conception of the present invention, it is also possible to base on
requirements to employ
any other possible weight calculation mode or calculation model in the prior
art, to which
no particular restriction is made in the embodiments of the present invention.
[0080] Fig. 3 illustrates a preferred mode of execution for the setup of
dimension features of the
uploader comprehensive score values and score calculation thereof As shown in
Fig. 3,
the three big dimension indicators of activity of the uploaders, quality of
the uploaders,
CA 03150500 2022-3-8

and verticality of the uploaders are mainly considered in this preferred mode.
[0081] The activities of the uploaders are one of the indicators, in which the
numbers of released
videos (combined with time decay) within the recent three months are employed
for
sorting, the sorting is mapped to a range of [x1,1], and it is avoided that
some few
uploaders differ from the majority in the order of magnitude, so as to render
the dimension
meaningless, wherein xl is a decimal between 0 and 1, and can be obtained
through model
learning as a weight parameter. Another indicator of activity is the volumes
of videos
played backed within the recent three months sorted according to time decay,
and mapped
to [x2,1], the aforementioned two indicators are multiplied to obtain scores
of the
dimension of activity of the uploaders. The quality scores of the uploaders
are calculated
by averaging the scores of such dimensions as sharing, praising, commenting,
the
proportion of positive comments, listing as favorites and following, and the
averaging
result is then multiplied with released video playback integrity rates to
obtain the scores
of the quality dimension. The verticality scores of the uploaders are obtained
by
calculating verticalities of category proportions of released videos within
the recent three
months. Finally, the three dimensions of the uploaders in terms of activity,
quality and
verticality are multiplied to obtain the final comprehensive score values of
the uploaders.
[0082] As should be noted, the evaluation of the above sorting N and the
specific time period
selection for the second preset period can be correspondingly set according to

requirements, for instance, N can be set as 3 and the second preset period can
be set as
three months, to which no particular restriction is made in the embodiments of
the present
invention.
[0083] As is notable, the process of step 101 can as well be realized by modes
other than the
mode recited in the aforementioned step, and these specific modes are not
restricted in
the embodiments of the present invention.
16
CA 03150500 2022-3-8

[0084] 102 - making statistics on the released video data of the target
uploaders according to a
preset tagging rule, and generating corresponding uploader word vector of one
or more
video category tag(s).
[0085] The specific categories and number of the video category tags can be
correspondingly set
according to requirements, for instance, the video category can include
athletics, finance
and economics, and amusement, etc., to which no particular restriction is made
in the
embodiments of the present invention.
[0086] 103 - obtaining user played-back video data within a first preset
period, making statistics
on the user played-back video data according to the preset tagging rule, and
generating
corresponding user word vector of the one or more video category tag(s).
[0087] Specifically, hotspot videos and miss-clicked videos are eliminated
from the user played-
back video data, top N target user tags each of whose number of videos
occupies a
proportion that is not lower than a preset proportion are counted out
according to the
preset tagging rule, the target user tags are calculated, and corresponding
user word
vectors of the target user tags are generated, wherein N is an integer greater
than 1.
[0088] As is notable, the process of step 103 can as well be realized by modes
other than the
mode recited in the aforementioned step, and these specific modes are not
restricted in
the embodiments of the present invention.
[0089] 104 - correspondingly matching the corresponding video category tags of
the uploader
word vector and the user word vector, obtaining an uploader word vector result
that has
reached a target matching degree with the user, and determining corresponding
uploader
information according to the uploader word vector result.
[0090] Fig. 4 illustrates a preferred mode of execution for the generation of
the uploader word
17
CA 03150500 2022-3-8

vector and the user word vector, and the corresponding tags matching. As shown
in Fig.
4, in this preferred mode, the uploaders employ the word vector representation
mode
calculated and obtained according to videos released in the recent three
months in
combination with time delay. Vector representations of users are obtained
according to
playback history in the recent one month, top 3 with each category occupying a
proportion
greater than 10% are counted in accordance with video classification tags of
watching
history, it is found out through statistical analysis that these are indeed
main interest points
of the users, while other videos in the watching history are usually some
hotspot videos
or miss-clicked cases that should be eliminated, so as to ensure precision of
user portraits.
[0091] As is notable, the process of step 104 can as well be realized by modes
other than the
mode recited in the aforementioned step, and these specific modes are not
restricted in
the embodiments of the present invention.
[0092] In addition, preferably, besides comprising the aforementioned steps
101, 102, 103 and
104, the uploader matching method provided by the embodiments of the present
invention
further comprises the following step:
[0093] recommending the uploader information to the user; and/or pushing a
video of the video
category tag corresponding to the uploader word vector result to the user.
Specifically,
top ranking Ns are selected through similarity calculation between vectors of
different
interest tags of users and the uploaders, an uploader list is obtained
according to
proportions under different interest tags of the users, and the list serves as
a candidate set
for recommendation to target users after merging and duplicate-removal,
wherein N is an
integer greater than 1, and the specific evaluation thereof can be set
according to
requirements.
[0094] Fig. 5 is a view schematically illustrating the structure of the
uploader matching device
provided by the embodiments of the present invention. As shown in Fig. 5, the
uploader
matching device 2 provided by the embodiments of the present invention mainly
18
CA 03150500 2022-3-8

comprises a calculating module 21, an uploader word vector generating module
22, a user
word vector generating module 23 and a matching module 24.
[0095] The calculating module 21 is employed for obtaining released video data
of uploaders,
determining comprehensive score values of the uploaders from one or more
dimension
feature score(s) according to the released video data, and screening out
target uploaders
according to the comprehensive score values of the uploaders.
[0096] Specifically, the calculating module 21 includes a first calculating
sub-module 211, a
second calculating sub-module 212 and a screening sub-module 213, of which the
first
calculating sub-module 211 is employed for calculating score(s) of one or more

dimension feature(s) in released video activity scores of the uploaders, video
quality
scores of the uploaders, and video verticality scores of the uploaders
according to the
released video data; the second calculating sub-module 212 is employed for
calculating
the comprehensive score values of the uploaders according to the one or more
dimension
feature score(s); and the screening sub-module 213 is employed for selecting
the
uploaders who rank top N to serve as the target uploaders according to a
sequence of the
comprehensive score values of the uploaders arranged in a decreasing order,
wherein N
is an integer greater than 1.
[0097] Preferably, the first calculating sub-module 211 is employed for:
sorting the number of
released videos of the uploaders within a second preset period, and a volume
of released
videos played back within the second preset period respectively in combination
with time
decay, mapping the sorted number of released videos and the sorted volume of
released
videos played back within the second preset period to a range of' [x1,1] and a
range of
[x2,1], determining respective weight indices of the number of released videos
and the
volume of released videos played back, wherein a first weight index xl and a
second
weight index x2 are each evaluated as a decimal between 0 and 1, thereafter
multiplying
the respective weight indices of the number of released videos with the
respective weight
19
CA 03150500 2022-3-8

indices of the volume of released videos played back, and calculating to
obtain the
released video activity scores of the uploaders;
[0098] sorting the number of sharings, the number of praisings, the number of
commentings, a
proportion of positive comments, the number of listings as favorites, the
number of
followings and released video playback integrity rates of released videos of
the uploaders
within the second preset period respectively combined with time decay, mapping
the
sorted number of sharings, number of praisings, number of commentings,
proportion of
positive comments, number of listings as favorites, number of followings and
released
video playback integrity rates to a range of [x3,1], a range of [x4,1], a
range of [x5,1], a
range of [x6,1], a range of [x7,1], a range of [x8,1] and a range of [x9,1],
respectively
determining respective weight indices of the number of sharings, the number of
praisings,
the number of commentings, the proportion of positive comments, the number of
listings
as favorites, the number of followings and the released video playback
integrity rates,
wherein a third weight index x3, a fourth weight index x4, a fifth weight
index x5, a sixth
weight index x6, a seventh weight index x7, an eighth weight index x8 and a
ninth weight
index x9 are each evaluated as a decimal between 0 and 1, thereafter summating
and
averaging the respective weight indices of the number of sharings, the number
of
praisings, the number of commentings, the proportion of positive comments, the
number
of listings as favorites and the number of followings, subsequently
multiplying the
summating and averaging result with the respective weight indices of the
released video
playback integrity rates, and calculating to obtain the video quality scores
of the uploaders;
and sorting category proportions of released videos of the uploaders within
the second
preset period in combination with time decay, mapping the sorted category
proportions
to a range of [xl 0,11, determining respective weight indices of the category
proportions,
wherein a tenth weight index x10 is evaluated as a decimal between 0 and 1,
thereafter
multiplying the respective weight indices of the category proportions, and
calculating to
obtain the video verticality scores of the uploaders.
[0099] The step of obtaining the first weight index xl, the second weight
index x2, the third
CA 03150500 2022-3-8

weight index x3, the fourth weight index x4, the fifth weight index x5, the
sixth weight
index x6, the seventh weight index x7, the eighth weight index x8, the ninth
weight index
x9 and the tenth weight index x10 is as follows:
[0100] respectively taking various dimension feature scores to which the first
weight index xl,
the second weight index x2, the third weight index x3, the fourth weight index
x4, the
fifth weight index x5, the sixth weight index x6, the seventh weight index x7,
the eighth
weight index x8, the ninth weight index x9 and the tenth weight index x10
correspond as
independent variables, taking following degrees of the uploaders after
exposure as
dependent variables, and employing a RandomForest algorithm and a GBDT
algorithm
to calculate the first weight index xl, the second weight index x2, the third
weight index
x3, the fourth weight index x4, the fifth weight index x5, the sixth weight
index x6, the
seventh weight index x7, the eighth weight index x8, the ninth weight index x9
and the
tenth weight index x10.
[0101] The second calculating sub-module 212 is employed for: multiplying the
released video
activity scores, the video quality scores and the video verticality scores,
and calculating
to obtain the comprehensive score values of the uploaders.
[0102] The uploader word vector generating module 22 is employed for making
statistics on the
released video data of the target uploaders according to a preset tagging
rule, and
generating corresponding uploader word vector of one or more video category
tag(s).
[0103] The user word vector generating module 23 is employed for obtaining
user played-back
video data within a first preset period, making statistics on the user played-
back video
data according to the preset tagging rule, and generating corresponding user
word vector
of the one or more video category tag(s). Specifically, hotspot videos and
miss-clicked
videos are eliminated from the user played-back video data, top N target user
tags each
of whose number of videos occupies a proportion that is not lower than a
preset proportion
are counted out according to the preset tagging rule, the target user tags are
calculated,
21
CA 03150500 2022-3-8

and corresponding user word vectors of the target user tags are generated,
wherein N is
an integer greater than 1.
[0104] The matching module 24 is employed for correspondingly matching the
corresponding
video category tags of the uploader word vector and the user word vector,
obtaining an
uploader word vector result that has reached a target matching degree with the
user, and
determining corresponding uploader information according to the uploader word
vector
result.
[0105] Preferably, the uploader matching device further comprises a data
recommending module
25 for recommending the uploader information to the user; and/or pushing a
video of the
video category tag corresponding to the uploader word vector result to the
user.
[0106] As should be noted, when the uploader matching device provided by this
embodiment
triggers uploader matching, the division into the aforementioned various
functional
modules is merely by way of example, while it is possible, in actual
application, to base
on requirements to assign the functions to different functional modules for
completion,
that is to say, to divide the internal structure of the device into different
functional
modules to complete the entire or partial functions described above. In
addition, the
uploader matching device provided by this embodiment pertains to the same
conception
as the uploader matching method provided by the method embodiment ¨ see the
corresponding method embodiment for its specific realization process, while no
repetition
will be made in this context.
[0107] A preferred mode for the uploader matching method and device provided
by the
embodiments of the present invention to perform uploader matching business is
introduced below. In this preferred mode of execution, a word segmentation
tool carries
a lexicon therewith, entertainment stars, film and TV drama names, sports
stars and team
information are additionally added as supplementary lexicon, Netease news,
Baidu
22
CA 03150500 2022-3-8

encyclopedia and Wikipedia obtained by a crawler system constitute a massive
corpus,
and word segmentation and word vector training are performed with respect to
the corpus
to finally obtain word vector representation of each word, dimensions of the
word vectors
are 200 dimensions, as determined by test effect, and normalization is then
performed on
the vectors.
[0108] Under the aforementioned corpus, TF-IDF training is carried out to
obtain IDF values,
normalization is then performed, weight enhancement is subsequently performed
on the
supplementary lexicon as 1, similar to the attention mechanism, more attention
is paid to
these words.
[0109] See the following Table 1 for video information, in which are carried
video id, video title
information, classification tags, video tag information, and releasing times,
etc. The video
information is word-segmented, a word vector table of words is searched, and
the word
vector representation of the current video is obtained in combination with
weighted
calculation of an IDF value table (performing normalization).
Table 1 Video Information Table
id title cata_name tag_names release
time
3714869 Wu lei has athletics athletics, 2019-08-
01
become the European 10:07:32
Mr. Key, Championship,
winning odds Wu lei, sports
show Espanyol lottery, lottery
will bring information
surprise in the station, sports
new season lottery video
collection
23
CA 03150500 2022-3-8

[0110] Drawing a portrait of a user is a process of calculating user word
vector, and the target
user group directed is a group of active users, namely recently active users
(having
playback records within 7 days lately) with certain volume of playbacks
(exceeding 10
videos) within a period lately (such as the recent 30 days). Word vector
calculation of the
users is particularized according to tag categories, for instance, a user
played 100 videos
within a period lately, of which there are 60 relevant to athletics, 20
relevant to finance
and economics, 15 relevant to amusement, 4 relevant to the society, and 1
relevant to
health; during the process of user portrait, a portrait is drawn for the user
under tag
categories ranking top 3 in the proportions and with the proportions each
exceeding 10%,
through which method it is made possible to obtain the main interest points of
the user,
and to remove few miss-clicked operations and hotspot videos that cannot
represent the
interest points of the user. In this example, athletics occupies 60%, finance
and economics
occupy 20%, amusement occupies 15%, society occupies 4%, and health occupies
1%,
so the portrait is drawn for the user in terms of the three dimensions of
athletics, finance
and economics, and amusement with respect to the current user, and word vector

representations of the corresponding dimensions of the user are calculated.
[0111] During the process of calculating user word vectors under different tag
categories of the
user, the time decay factor (for instance, the decay period is 5 days, the
decay coefficient
is 0.95, taking for example a video played back 12 days before the current
date, the video
crosses two decay periods, and should be decayed by 0.951'2) is combined to
calculate
the word vector representations of the user.
[0112] The three big dimensions of activity of the uploaders, quality of the
uploaders, and
verticality of the uploaders are comprehensively considered to facilitate
subsequent
comprehensive scoring of the qualities of the uploaders.
[0113] The activity indicators of the uploaders are sorted by a combination of
the number of
videos released in the recent three months with time decay, the sorting result
is mapped
24
CA 03150500 2022-3-8

to a range of [x1,1] (it is avoided that some few uploaders differ from the
majority in the
order of magnitude, so as to render the dimension meaningless), wherein xl is
a decimal
between 0 and 1, and is obtained through model learning as a model parameter.
Another
indicator of activity is the volumes of videos played backed within the recent
three
months sorted according to time decay, and mapped to [x2,1]. The
aforementioned two
indicators are multiplied to obtain scores of the dimension of activity of the
uploaders.
[0114] The quality scores of the uploaders are calculated by averaging the
scores of such
dimensions as sharing, praising, commenting, the proportion of positive
comments,
listing as favorites and following (that are respectively sorted and
thereafter mapped to
between a certain variable x and 1, wherein x is a decimal between 0 and 1),
and the
averaging result is then multiplied with released video playback integrity
rates to obtain
the scores of the quality dimension.
[0115] The verticality scores of the uploaders are obtained by calculating
verticalities of category
proportions of released videos within the recent three months, sorted and
mapped to
between a certain variable x and 1, wherein xis a decimal between 0 and 1.
[0116] While the various dimension scores are calculated, different importance
degrees of the
dimensions in the calculation of the comprehensive scores are decided by
mapping them
to the corresponding scoring ranges, namely the weights of the various
dimension scores,
these parameters are obtained through model training, as shown in the
following Table 2.
Table 2 Mapping Datasheet of Various Dimension Scores
Parameter xl x2 x3 x4 x5
CA 03150500 2022-3-8

Meaning sorting of numbers sorting of sorting of
sorting of sorting of
of videos released in volumes of numbers of
numbers of numbers of
the recent three released videos sharings, mapped
prai sings, commentings,
months, mapped to integrally played to [x3,1]
mapped to [x4,1] mapped to [x5,1]
[x1,1] back in the recent
three months to
before three days,
mapped to [x2,1]
Range xl is a decimal x2 is a decimal x3 is
a decimal x4 is a decimal x5 is a decimal
between 0 and 1 between 0 and 1 between 0 and 1
between 0 and 1 between 0 and 1
Parameter x6 x7 x8 x9 x10
Meaning sorting of sorting of sorting of
summarized verticality scoring
proportions of numbers of numbers of
sorting of of various types
positive comments, listings as followings,
released video of uploaders
mapped to [x6,1] favorites, mapped mapped to [x8,11
playback (sorting of various
to [x7,1] integrity rates
in category
the recent three
proportions of
months to before videos released in
three days, recent
three
mapped to [x9,1] months, mapped
to [x10,1])
Range x6 is a decimal x7 is a decimal x8 is
a decimal x9 is a decimal x10 is a decimal
between 0 and 1 between 0 and 1 between 0 and 1
between 0 and 1 between 0 and 1
[0117] The various dimension features constructed above serve as independent
variables, the
following degrees of the uploaders (uploaders exposed to users being followed
by users)
serve as dependent variables, and the weights of the aforementioned indicators
are
obtained through modeling and training.
26
CA 03150500 2022-3-8

[0118] The comprehensive scores of the qualities of uploaders are obtained by
multiplying the
three dimension scores in terms of activity indicator, quality and verticality
of the
uploaders, and these scores are sorted and mapped to between [0,1000] to serve
as the
comprehensive scores of the qualities of the uploaders. While uploaders are
being
recommended to users, only the uploaders with higher qualities are selected
(for example,
uploaders ranking top 600 in the comprehensive scores of qualities are
selected).
[0119] The calculation of uploader word vectors (with respect to high-quality
uploaders) is based
on videos released by uploaders within the recent three months, and is similar
to the
process of calculating user-particularized word vectors, whereby top 3 in the
proportions
with each category occupying a proportion greater than 10% are selected
according to
classification tags of released videos of the uploaders to finely calculate
word vector
representations (word vector representations according to videos, calculated
and obtained
in combination with the time decay factor) of the uploaders in a plurality of
dimensions.
[0120] Similarities between users and uploaders are calculated through cosine
similarity
calculation; with respect to the scenario in which a target user is drawn with
a portrait
having a plurality of tagged dimensions, her/his result set of recommendation
of
uploaders is constituted according to proportions of the tags of the user, and
inversion is
finally made according to similarities for recommendation to the user.
[0121] All the aforementioned optional technical solutions are randomly
combinable to form
optional embodiments of the present invention, to which no repetition is made
on a one-
by-one basis.
[0122] To sum it up, in comparison with prior-art technology, the uploader
matching method and
device provided by the embodiments of the present invention achieve the
following
advantageous effects.
27
CA 03150500 2022-3-8

[0123] 1. By sorting out in overall such data as released video information of
uploaders and user
behaviors, history records of user playbacks and such multi-dimensional
information as
relevant to the listing as favorites, sharing, praising, commenting and
playback integrities
of videos released by uploaders are obtained, a comprehensive scoring scheme
based on
multiple data sources and multi-dimensional variables achieves comprehensive
evaluation of uploaders, and high-quality uploader information is extracted
therefrom.
[0124] 2. Various dimension weights for evaluating uploader qualities are
obtained through
model training by corresponding algorithms, and evaluation of uploader
qualities is made
more precise.
[0125] 3. Uploaders are finely classified according to released video category
tags, whereby
precision of uploader portraits is enhanced.
[0126] 4. During the process of drawing portraits of users, hotspot videos and
possibly miss-
clicked videos are eliminated, portraits are drawn for the users by
differentiating different
interest tags, result sets are proportionally formed, and precision in
recommending result
sets is enhanced.
[0127] 5. Time decay is taken into consideration in the processes of
calculating user vector and
calculating uploader vector, whereby transfer of interest points is enhanced.
[0128] 6. By matching the high-quality uploader word vector with the user word
vector of a
precise user portrait, high-quality uploaders are finally recommended to
target users with
high matching degree, by following uploaders similar to themselves, users are
facilitated
to timely watch high-quality videos, whereby satisfaction of the users is
enhanced,
comparison report indicators are obtained through AB test, whereby CTR, user
playback
volume and average playback integrity are enhanced, and user experience is
enhanced as
a whole.
28
CA 03150500 2022-3-8

[0129] As understandable by persons ordinarily skilled in the art, realization
of the entire or
partial steps of the aforementioned embodiments can be completed by hardware,
or by a
program instructing relevant hardware, the program can be stored in a computer-
readable
storage medium, and the storage medium can be a read-only memory, a magnetic
disk, or
an optical disk, etc.
[0130] The embodiments of the present application are described with reference
to flowcharts
and/or block diagrams of the method, device (system), and computer program
product
embodied in the embodiments of the present application. As should be
understood, each
flow and/or block in the flowcharts and/or block diagrams, and any combination
of flow
and/or block in the flowcharts and/or block diagrams can be realized by
computer
program instructions. These computer program instructions can be supplied to a
general
computer, a dedicated computer, an embedded processor or a processor of any
other
programmable data processing device to form a machine, so that the
instructions executed
by the computer or the processor of any other programmable data processing
device
generate a device for realizing the functions designated in one or more
flow(s) of the
flowcharts and/or one or more block(s) of the block diagrams.
[0131] These computer program instructions can also be stored in a computer-
readable memory
enabling a computer or any other programmable data processing device to
operate by a
specific mode, so that the instructions stored in the computer-readable memory
generate
a product containing instructing means, and this instructing means realizes
the functions
designated in one or more flow(s) of the flowcharts and/or one or more
block(s) of the
block diagrams.
[0132] These computer program instructions can also be loaded onto a computer
or any other
programmable data processing device, so as to execute a series of operations
and steps
on the computer or the any other programmable device to generate computer-
realized
29
CA 03150500 2022-3-8

processing, so that the instructions executed on the computer or the any other

programmable device provide steps for realizing the functions designated in
one or more
flow(s) of the flowcharts and/or one or more block(s) of the block diagrams.
[0133] Although preferred embodiments in the embodiments of the present
application have
been described so far, it is still possible for persons skilled in the art to
make additional
modifications and amendments to these embodiments upon learning of the basic
inventive
conception. Accordingly, the attached Claims are meant to cover the preferred
embodiments and all modifications and amendments that fall within the scope of
the
embodiments of the present application.
[0134] Apparently, persons skilled in the art can make various amendments and
modifications to
the present invention without departing from the spirit and scope of the
present invention.
Thus, if such amendments and modifications to the present invention fall
within the
Claims of the present invention and equivalent technology, the present
invention is also
meant to cover these amendments and modifications.
[0135] What is described above is merely directed to preferred embodiments of
the present
invention, and they are not meant to restrict the present invention. Any
amendment,
equivalent replacement and improvement makeable within the spirit and scope of
the
present invention shall all be covered within the protection scope of the
present invention.
CA 03150500 2022-3-8

Representative Drawing