Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 03165616 2022-06-22
SPORTS NEWS WRITING METHOD BASED ON NATURAL LANGUAGE, DEVICE
AND ELECTRONIC EQUIPMENT
BACKGROUND OF THE INVENTION
Technical Field
[0001] The present invention relates to the field of natural language
processing technology, and
more particularly to a sports news writing method based on a natural language,
and
corresponding device and electronic equipment.
Description of Related Art
[0002] When the currently available template traversing and matching strategy
matches event
data and templates, it is required to compare slots with the numbers of slots
in the
templates on a one-by-one basis, until the proper template is found out. For
instance, a
piece of data of a goal event reads: {ORG NEU: Newcastle, PER ACT: Schaer,
EVEINF LOC FROM: center outside the penalty area, EVEINF BODY: right foot,
EVEINF LOC TO: right upper corner of the goal} ¨ this piece of data contains
five slots,
when templates are matched, it is supposed there are the following templates
for goal
events:
[0003] {ORG NEU} Get a goal! {PER-ACT} shoot, the ball flies from {EVEINF LOC
TO}
into the goal
[0004] {ORG NEU} Get a goal! {PER-ACT} {EVEINF LOC FROM} {EVEINF BODY}
shoot, the ball draws a perfect arc and flies from {EVEINF LOC TO} into the
goal.
[0005] Each template is traversed, a slot set contained in the current
template and information
concerning the number of times for which each slot appears are calculated
according to
regular matching, if the slot set of the current template is a subset of a
slot set in the data,
1
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
a template is successfully matched. As can be seen, it is required for the
traversing
strategy to calculate slot information of the templates each time, and then to
perform set
operation with slot information of the data, so it is relatively time-
consuming; because
there is performance requirement on the online system, return is usually
effected once
matching is successful, and the entire templates are not matched. This brings
about
another problem, i.e., the first successfully matched template that conforms
to the
condition is usually not the optimum template, that is to say, types of slots
and the number
of slots do not satisfy the maximum requirement. Taking the above for example,
successful matching is counted when the first template is matched, but use of
the first
template would prevent the EVEINF LOC FROM and the EVEINF BODY information
from being filled, so that information volume of the article as generated is
relatively small.
As a consequence, the traversing and matching strategy is not only low in
efficiency, but
is also inferior in terms of diversity of the matching results.
SUMMARY OF THE INVENTION
[0006] In view of the deficiencies prevalent in the state of the art, one of
the objectives of the
present application is to provide a sports news writing method based on a
natural language,
to enhance the diversity of sentence patterns of articles and maximize
information volume
of articles. The method comprises the following steps:
[0007] obtaining a corpus to be processed, an event set, slots, and a slot
value to which each slot
corresponds;
[0008] marking event templates in the corpus according to each event in the
event set, the slots,
and the slot values;
[0009] performing weight assignment on each event;
[0010] coding each event, and types and the number of the slots in the event
templates;
[0011] screening the events and the event templates according to the weight of
each event;
[0012] matching and filling the screened events and event templates, and
generating news
content; and
2
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
[0013] reprocessing the news content to obtain the final news content.
[0014] Preferably, the obtaining an event set, slots, and a slot value to
which each slot
corresponds includes the following steps:
[0015] obtaining a preset number of sports news corpora;
[0016] processing the sports news corpora to obtain all the events, slots, and
the slot value to
which each slot corresponds; and
[0017] disposing all the events in the same and single set to obtain the event
set.
[0018] Preferably, after processing the sports news corpora to obtain all the
events, slots, and the
slot value to which each slot corresponds, the method further comprises the
following
steps:
[0019] judging whether each event, each slot, and each slot value conform to a
preset range;
[0020] if yes, retaining the event, the slot, and the slot value;
[0021] if not, deleting the event, the slot, or the slot value.
[0022] Preferably, the event includes a title, an abstract, and a text.
[0023] Preferably, the performing weight assignment on each event includes the
following steps:
[0024] dividing the corpus according to all the events to obtain plural
sections;
[0025] constructing, with respect to each event, mapping between the event and
each section;
and
[0026] setting, with respect to each mapping, a weight of the event to which
the mapping
corresponds.
[0027] Preferably, the coding each event, and types and the number of the
slots in the event
templates includes the following steps:
[0028] obtaining the event templates and the events to be coded;
[0029] counting the event templates, and the types and the number of the slots
in the events
3
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
according to regular matching;
[0030] determining a total number m of all the slots and the maximum number of
times n for
which the slots appear in each event template;
[0031] determining that each slot should be assigned with n number of binary
digits for
representation according to the maximum number of times for which the slots
appear,
wherein n is a divisor of 64;
[0032] determining long types of a coding type as employed and the number x of
codes according
to the total number of the slots and the number of binary digits assigned to
each slot,
wherein x=[(m*n)/641+1;
[0033] traversing each slot in the event templates, and performing binary
coding on the number
of slots of the current slot;
[0034] determining that the current slot is coded at the yth long type
according to an index address
i of the current slot, wherein y=i/(64/n)+1;
[0035] moving binary representation of the number of slots leftwards for p
times, wherein p=(i-
(y-1)*(64/n)rn; and
[0036] joining all codes of the long type to obtain the final codes.
[0037] Preferably, the screening the events includes the following steps:
[0038] obtaining a corresponding weight of each event;
[0039] comparing each weight with a preset threshold; and
[0040] retaining any event whose weight is greater than the preset threshold,
and eliminating the
remaining events.
[0041] Preferably, the screening the event templates includes the following
steps:
[0042] obtaining the screened events and the codes corresponding thereto, and
the codes to which
all the event templates correspond in the events;
[0043] selecting one or more event template(s) with the maximum number of
slots to serve as
candidate event template(s); and
[0044] randomly selecting one event template from the candidate event
template(s) to serve as
4
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
the event template to be filled.
[0045] In view of the deficiencies prevalent in the state of the art, the
second objective of the
present application is to provide a sports news writing device based on a
natural language,
to enhance the diversity of sentence patterns of articles and maximize
information volume
of articles. The device comprises:
[0046] an obtaining unit, for obtaining a corpus to be processed, an event
set, slots, and a slot
value to which each slot corresponds;
[0047] an event template marking unit, for marking event templates in the
corpus according to
each event in the event set, the slots, and the slot values;
[0048] a weight assigning unit, for performing weight assignment on each
event;
[0049] a coding unit, for coding each event, and types and the number of the
slots in the event
templates;
[0050] a screening unit, for screening the events and the event templates
according to the weight
of each event;
[0051] a news content generating unit, for matching and filling the screened
events and event
templates, and generating news content; and
[0052] a news content processing unit, for reprocessing the news content to
obtain the final news
content.
[0053] In view of the deficiencies prevalent in the state of the art, the
third objective of the
present application is to provide an electronic equipment to enhance the
diversity of
sentence patterns of articles and maximize information volume of articles. The
electronic
equipment comprises:
[0054] at least one processor; and
[0055] a memory, communicably connected with the at least one processor;
wherein
[0056] the memory stores an instruction executable by the at least one
processor, and the
instruction is executed by the at least one processor to enable the processor
to execute the
aforementioned sports news writing method.
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
[0057] The sports news writing method, device and electronic equipment based
on a natural
language as provided by the present application make it possible to
automatically
generate news contents according to event templates extracted by prior
analysis of great
quantities of sports news and according to weight assignments of users
themselves on
events, whereby diversity of sentence patterns of articles is enhanced and
information
volume of articles is maximized on the one hand, highly effectively automatic
writing of
sports news articles is realized and the input of manpower cost is reduced on
the other
hand.
BRIEF DESCRIPTION OF THE DRAWINGS
[0058] To more clearly describe the embodiments of the present invention or
technical solutions
in the prior-art technology, drawings required to be used in the description
of the
embodiments or the prior-art technology are briefly introduced below.
Apparently, the
drawings as introduced below are merely directed to some embodiments of the
present
invention, and it is further possible for persons ordinarily skilled in the
art to acquire other
drawings based on these drawings without spending any creative effort in the
process.
[0059] Fig. 1 is a flowchart illustrating a sports news writing method based
on a natural language
provided by Embodiment 1 of the present invention;
[0060] Fig. 2 is a flowchart illustrating a sports news writing method based
on a natural language
provided by Embodiment 2 of the present invention;
[0061] Fig. 3 is a flowchart illustrating a sports news writing method based
on a natural language
provided by Embodiment 3 of the present invention;
[0062] Fig. 4 is a flowchart illustrating a sports news writing method based
on a natural language
6
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
provided by Embodiment 4 of the present invention;
[0063] Fig. 5 is a flowchart illustrating a sports news writing method based
on a natural language
provided by Embodiment 5 of the present invention;
[0064] Fig. 6 is a flowchart illustrating a sports news writing method based
on a natural language
provided by Embodiment 6 of the present invention;
[0065] Fig. 7 is a flowchart illustrating a sports news writing method based
on a natural language
provided by Embodiment 7 of the present invention;
[0066] Fig. 8 is a view schematically illustrating the structure of a sports
news writing device
based on a natural language provided by the present invention; and
[0067] Fig. 9 is a view schematically illustrating the structure of an
electronic equipment
provided by the present invention.
DETAILED DESCRIPTION OF THE INVENTION
[0068] The embodiments of the current disclosure are described in greater
detail below with
reference to the accompanying drawings.
[0069] The modes of execution of the current disclosure are described below
through specific
concrete examples, and it is possible for persons skilled in the art to easily
learn of other
advantages and effects of the current disclosure from the contents disclosed
in this
Description. Apparently, the embodiments as described are merely partial,
rather than the
entire, embodiments of the current disclosure. The current disclosure can be
further
implemented or applied through additional, different, specific modes of
execution, and
the various details in this Description can also be variously modified or
changed based
7
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
on different viewpoints and applications without departing from the spirits of
the current
disclosure. As should be noted, the following embodiments and features in the
embodiments can be combined with one another provided that they are not
contradictory
to one another. Based on the embodiments in the current disclosure, all other
embodiments obtainable by persons ordinarily skilled in the art without
spending any
creative effort should all be covered within the protection scope of the
current disclosure.
[0070] As should be noted, various aspects of the embodiments are described
below within the
range of the attached Claims. As is obvious, the aspects described in this
Description can
be embodied in a wide range of forms, and any specific structure and/or
function
described in this Description are/is merely explanative in nature. As should
be clear to
persons skilled in the art on the basis of the current disclosure, one aspect
described in
this Description can be implemented independently of any other aspect, and can
also be
combined with two or more aspects in various modes. By way of example, any
number
of aspects enunciated in this Description can be employed to implement the
device/equipment and/or to practice the method. In addition, other structures
and/or
functions than one or more aspects enunciated in this Description can be
employed to
implement the device/equipment and/or to practice the method.
[0071] As should be further noted, the illustrations provided in the following
embodiments
merely illustrate the basic conception of the current disclosure in an
illustrative mode,
shown in the illustrations are merely the components relevant to the current
disclosure
and these are not drawn in the number, shapes and sizes of the actual
components during
actual implementation. The forms, number and proportions of the various
components
are subjected to random changes during actual implementation, and the layouts
and forms
of the components might be even more complicated during actual implementation.
[0072] Moreover, specific details are provided in the following description to
facilitate thorough
understanding of the concrete examples. However, as understood by persons
skilled in
8
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
the art, the said aspects can still be practiced in the absence of these
specific details.
[0073] An embodiment of the current disclosure provides a sports news writing
method based
on a natural language. The sports news writing method based on a natural
language
provided by this embodiment can be executed by a computer system, the computer
system
can be embodied as software, or a combination of software with hardware, and
the
computer system can be integrated in a server, a terminal equipment, etc.
[0074] Embodiment 1
[0075] As shown in Fig. 1, in an embodiment of the present application is
provided a sports news
writing method based on a natural language, and the method comprises the
following
steps.
[0076] Step S101 - obtaining a corpus to be processed, an event set, slots,
and a slot value to
which each slot corresponds.
[0077] In this step, there are many methods of obtaining a corpus to be
processed, an event set,
slots, and a slot value to which each slot corresponds, for instance, it is
possible to
manually input preset corpus to be processed, event set, slots, and slot value
to which
each slot corresponds to an equipment, it is also possible for the equipment
to
automatically obtain preset corpus to be processed, event set, slots, and slot
value to
which each slot corresponds, and it is further possible for a server to
automatically crawl
(by using a crawler for example) preset corpus to be processed, event set,
slots, and slot
value to which each slot corresponds, to which modes the present application
makes no
restriction.
[0078] Step S102 - marking event templates in the corpus according to each
event in the event
set, the slots, and the slot values.
9
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
[0079] In this step, in accordance with each event in the event set, the
slots, and the slot values
obtained in step S101, it is possible to mark event templates in the corpus to
be processed
as obtained in step S101.
[0080] Step S103 - performing weight assignment on each event.
[0081] In this step, there are many methods of performing weight assignment on
each event, for
instance, it is possible to manually input a weight to which each event
corresponds to an
equipment, it is also possible for the equipment to automatically obtain a
weight to which
each event corresponds, and it is further possible for a server to
automatically crawl (by
using a crawler for example) a weight to which each event corresponds, to
which modes
the present application makes no restriction.
[0082] Step S104 - coding each event, and types and the number of the slots in
the event
templates.
[0083] In this step, there are many methods of coding each event, and types
and the number of
the slots in the event templates, for instance, it is possible to manually
input codes of each
event, and types and the number of the slots in the event templates to an
equipment, it is
also possible for the equipment to automatically obtain codes of each event,
and types
and the number of the slots in the event templates, and it is further possible
for a server
to automatically crawl (by using a crawler for example) codes of each event,
and types
and the number of the slots in the event templates, to which modes the present
application
makes no restriction.
[0084] In this step, the coding type employed can be binary, octonary or
decimal, to which the
present application makes no restriction.
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
[0085] Step S105 - screening the events and the event templates according to
the weight of each
event.
[0086] In this step, the events and the event templates are screened according
to the weight of
each event obtained in step S103. There are many screening standards, for
instance, it is
possible to compare the weight of each event with a preset threshold, and to
eliminate
any event whose weight is smaller than the preset threshold, while the present
application
makes no restriction to the screening standards.
[0087] Step S106 - matching and filling the screened events and event
templates, and generating
news content.
[0088] In this step, the events and event templates screened in step S105 are
matched and filled
to generate news content. There are many matching and filling modes, to which
no
restriction is made in the present application.
[0089] Step S107 - reprocessing the news content to obtain the final news
content.
[0090] In this step, the news content obtained in step S106 is reprocessed to
obtain the final news
content. There can be many reprocessing modes, such as manual reviewing and
retouching, or rechecking by the system, etc., to which no restriction is made
in the
present application.
[0091] In Embodiment 1, the sports news writing method based on a natural
language as
provided by the present application makes it possible to automatically process
the corpus
to be processed, so as to enable automatic generation of news contents, to
reduce
manpower, and to enhance automatic processing capability for news corpora.
[0092] Embodiment 2
11
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
[0093] As shown in Fig. 2, in an embodiment of the present application,
obtaining an event set,
slots, and a slot value to which each slot corresponds in step S101 includes
the following
steps.
[0094] Step S201 - obtaining a preset number of sports news corpora.
[0095] In this step, there are many methods of obtaining a preset number of
sports news corpora,
for instance, it is possible to manually input a preset number of sports news
corpora to an
equipment, it is also possible for the equipment to automatically obtain a
preset number
of sports news corpora, and it is further possible for a server to
automatically crawl (by
using a crawler for example) a preset number of sports news corpora, to which
modes the
present application makes no restriction.
[0096] In this step, the preset number can be theoretically one piece, and can
also be 100,000
pieces, however, out of overall considerations of data processing amount and
accuracy of
the result, the preset number can be selected as 1,000 pieces.
[0097] An example is taken for explanation by selecting two pieces of sports
news corpora,
specifically as follows:
[0098] sports news corpus 1: star player A shoots, goal succeeds.
[0099] sports news corpus 2: star player B shoots unsuccessfully, goal fails.
[0100] Step S202 - processing the sports news corpora to obtain all the
events, slots, and the slot
value to which each slot corresponds.
[0101] In this step, the sports news corpora in step S201 are processed to
obtain all the events,
slots, and the slot value to which each slot corresponds.
12
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
[0102] Specifically, an example is taken for explanation with the above two
pieces of sports news
corpora, specifically as follows:
[0103] event is: goal;
[0104] slot is: __ shoots (the underlined location is the slot);
[0105] slot values are: star player A shoots, star player B shoots (the
underlined locations are the
slot values, in other words, the slot values are star player A and star player
B).
[0106] Step S203 - disposing all the events in the same and single set to
obtain the event set.
[0107] In this step, all the events obtained in step S202 are disposed in the
same and single set
to obtain an event set. The event set contains at least one event, and the
number of event(s)
contained in the event set may differ according to classifying standards of
events. For
instance, the event can be selected as "goal" in step S202, then the event set
in step S203
contains only one event ¨ "goal"; it is also possible to select "goal
succeeds" and "goal
fails" as events in step S202, then the event set in step S203 contains two
events -- "goal
succeeds" and "goal fails".
[0108] In the embodiments of the present application, an event can further
include a title, an
abstract, and a text. For instance, a piece of sports news generally at least
includes the
three sections of "title", "abstract", and "text"; of course, it can further
include "subtitle",
"editorial note", and "commentary", etc., to which no restriction is made in
the present
application.
[0109] In Embodiment 2, the sports news writing method based on a natural
language as
provided by the present application makes it possible to analyze and process
great
quantities of sports news corpora to obtain therefrom event sets, slots, and
slot values to
which each slot corresponds, so as to provide research templates for
subsequent automatic
generation of news contents.
13
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
[0110] Embodiment 3
[0111] As shown in Fig. 3, in an embodiment of the present application, after
processing the
sports news corpora to obtain all the events, slots, and the slot value to
which each slot
corresponds in step S202, the following steps are further included:
[0112] Step S301 -judging whether each event, each slot, and each slot value
conform to a preset
range;
[0113] Step S302 - if yes, retaining the event, the slot, and the slot value;
[0114] Step S303 - if not, deleting the event, the slot, or the slot value.
[0115] Through the above steps S301-S303, each event, each slot, and each slot
value in step
S202 can be judged as to whether they conform to a preset range, and to
perform
eliminating or retaining operation.
[0116] Specifically, the preset range can be either a specific numerical value
or a definitive
statement. For instance, plural events can be selected from the 1,000 pieces
of sports news
as "shoot", "goal", and "penalty kick", of which "penalty kick" is
proportionally lower
relative to "shoot" and "goal", and the event of "penalty kick" can be
eliminated in the
case processing result precision is required not much highly, whereby data
processing
amount is greatly reduced.
[0117] In Embodiment 3, the sports news writing method based on a natural
language as
provided by the present application makes it possible to screen events, slots
and slot
values to which each slot corresponds as obtained, whereby accuracy of news
generation
is enhanced on the one hand, and data processing amount is reduced on the
other hand.
[0118] Embodiment 4
[0119] As shown in Fig. 4, in an embodiment of the present application,
performing weight
14
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
assignment on each event in step S103 includes the following steps:
[0120] Step S401 - dividing the corpus according to all the events to obtain
plural sections;
[0121] Step S402 - constructing, with respect to each event, mapping between
the event and each
section; and
[0122] Step S403 - setting, with respect to each mapping, a weight of the
event to which the
mapping corresponds.
[0123] Specific explanation is made below with Table 1 and Table 2:
[0124] Table 1
[0125]
Event Section First Paragraph Second Paragraph
Title 0.6 0.3
Abstract 0.4 0.7
[0126] With respect to Table 1, the corpus is divided into a first paragraph
and a second paragraph,
of which the first paragraph is considered by default to be the title section,
and the second
paragraph is considered by default to be the abstract section (in some news,
the abstract
may even be larger than one paragraph, but this case is not considered for the
sake of
brevity).
[0127] When weights are assigned to the title event and the abstract event,
the weights of the
title event with respect to the first paragraph and the second paragraph are
respectively
0.6 and 0.3, and the weights of the abstract event with respect to the first
paragraph and
the second paragraph are respectively 0.4 and 0.7, that is to say, when the
system
considers the importance of the title event and the abstract event in the
first paragraph
and the second paragraph of the corpus, it is apparent that the title event is
most probably
in the first paragraph, and that the abstract event is most probably in the
second paragraph.
Of course, in some other embodiments, when ranking is made according to a
descending
order of weights, the abstract event is most probably in the first paragraph,
while the title
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
event is most probably in the second paragraph.
[0128] Table 2
[0129]
Event Section First Paragraph Second Paragraph
Goal 0.6 0.7
Penalty Kick 0.4 0.3
[0130] With respect to Table 2, the corpus is divided into a first paragraph
and a second paragraph,
of which the first paragraph is considered by default to be the title section,
and the second
paragraph is considered by default to be the abstract section (in some news,
the abstract
may even be larger than one paragraph, but this case is not considered for the
sake of
brevity).
[0131] When weights are assigned to the goal event and the penalty kick event,
the weights of
the goal event with respect to the first paragraph and the second paragraph
are
respectively 0.6 and 0.7, and the weights of the penalty kick event with
respect to the first
paragraph and the second paragraph are respectively 0.4 and 0.3, that is to
say, when the
system considers the importance of the goal event and the penalty kick event
in the first
paragraph and the second paragraph of the corpus, it is apparent that the goal
event is
most probably in the first paragraph, and that the penalty kick event is most
probably in
the second paragraph. Of course, in some other embodiments, when ranking is
made
according to a descending order of weights, the penalty kick event is most
probably in
the first paragraph, while the goal event is most probably in the second
paragraph.
[0132] In Embodiment 4, the sports news writing method based on a natural
language as
provided by the present application makes it possible to assign weight to each
event, and
it is possible to set event weights by self-definition according to
requirements on the one
hand, and to also provide basis for subsequent coding operation on the other
hand.
16
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
[0133] Embodiment 5
[0134] As shown in Fig. 5, in an embodiment of the present application, coding
each event, and
types and the number of the slots in the event templates in step S104 includes
the
following steps.
[0135] Step S501 - obtaining event templates and events to be coded.
[0136] In this step, there are many methods of obtaining event templates and
events to be coded,
for instance, it is possible to manually input event templates and events to
be coded to an
equipment, it is also possible for the equipment to automatically obtain event
templates
and events to be coded, and it is further possible for a server to
automatically crawl (by
using a crawler for example) event templates and events to be coded, to which
modes the
present application makes no restriction.
[0137] In the embodiments of the present application, the coding as employed
is binary coding.
[0138] Step S502 - counting the event templates, and the types and the number
of the slots in the
events according to regular matching.
[0139] In this step, a regular expression is employed to sequentially traverse
the event templates
and events to be coded, so that it is possible to count the types and the
number of each
type of the event templates, and the types and the number of each type of the
slots in the
events.
[0140] Step S503 - determining a total number m of all the slots and the
maximum number of
times n for which the slots appear in each event template.
17
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
[0141] In this step, through step S502 it is possible to determine a total
number m of all the slots
and the maximum number of times n for which the slots appear in each event
template.
[0142] In the embodiments of the present application, since binary coding is
employed, so m=22,
n=4. The maximum 4 times for which slots appear in the event template are
based on to
determine that 4 bits of binary digits should be assigned to each event slot
to represent
the number of times for which the event slot appears, and in the 4 bits of
binary digits, a
1 on each binary digit represents that the event slot appears for one time.
[0143] Step S504 - determining that each slot should be assigned with n number
of binary digits
for representation according to the maximum number of times for which the
slots appear,
wherein n is a divisor of 64.
[0144] Step S505 - determining long types of a coding type as employed and the
number x of
codes according to the total number of the slots and the number of binary
digits assigned
to each slot, wherein x=[(m*n)/641+1.
[0145] In steps S504-S505, according to the total number 22 of the slots and
the 4 bits of binary
digits assigned to each slot, it is derived that 22*4=88 bits of binary digits
are required to
represent the number of all the slots.
[0146] In the embodiments of the present application, out of overall
considerations of reality and
the data processing amount, long type data are used for coding. One piece of
long type
data consists of 64 bits, so n should be a divisor of 64. At this time, 2
pieces of long type
data are required to code the event templates and the slots in the corpus, and
the 2 pieces
of long type data are both initialized as 0.
[0147] Step S506 - traversing each slot in the event templates, and performing
binary coding on
the number of slots of the current slot.
18
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
[0148] In this step, the long type is employed to sequentially traverse each
slot in the event
templates, and binary coding is performed on each slot number. For instance, a
certain
slot appears for two times in an event template, then it is binarily
represented by the long
type as 0011.
[0149] Step S507 - determining that the current slot is coded at the yth long
type according to an
index address i of the current slot, wherein y=i/(64/n)+1.
[0150] In this step, it is determined that the current slot is coded at the
yth long type according to
an index address i of the current slot, and the y value can be calculated by
the following
formula:
[0151] y=i/(64/4)+1,
[0152] where i is an index address of the current slot. For instance, the
index address i of the
current slot is 15, then y=1, expressing that the current slot is coded on the
first long data;
if the current slot i is 16, then y=2, the current slot should be coded on the
second long
data.
[0153] Step S508 - moving binary representation of the number of slots
leftwards for p times,
wherein p=0-(y-1)*(64/n))*n.
[0154] In this step, binary representation of the number of slots is moved
leftwards for p times,
and p therein is calculated as follows:
[0155] p=0-(y-1)*(64/4))*4
[0156] where i is a slot index address, and y represents that the current slot
is coded on the yth
long data. For instance, the index address i of a certain slot is 10, it is
then calculated to
derive that y=1, that is to say, the slot is coded on the first long data, and
the slot appears
for two times in the event template, then the number of times of the slot is
binarily
represented as 0011, during coding, 0011 is moved leftwards by 10*4 bits, to
derive the
19
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
following codes: 0000 0000 0000 0000 0000 0011 0000 0000 0000 0000 0000 0000
0000
0000 0000 0000.
[0157] Step S509 -joining all codes of the long type to obtain the final
codes.
[0158] In this step, all codes of the long type are joined, so as to obtain
the final codes. All slot
codes on the same and single long data should be summated.
[0159] In Embodiment 5, the sports news writing method based on a natural
language as
provided by the present application makes it possible to binarily code each
event and the
types and numbers of slots in the event templates, whereby the subsequent
operation of
screening the events and event templates is facilitated, and the workload of
the screening
operation is reduced.
[0160] Embodiment 6
[0161] As shown in Fig. 6, in an embodiment of the present application,
screening the events in
step S105 includes the following steps:
[0162] Step S601 - obtaining a corresponding weight of each event;
[0163] Step S602 - comparing each weight with a preset threshold; and
[0164] Step S603 - retaining any event whose weight is greater than the preset
threshold, and
eliminating the remaining events.
[0165] The many events should be screened in steps S601-S603, and the
screening standard is
the comparison of the weights corresponding to the events with a preset
threshold.
[0166] A specific example is taken below for explanation.
[0167] In the embodiments of the present application, the news includes three
sections, namely
a title, an abstract, and a text, and the events include two types, namely
"goal" and
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
"penalty kick". The two events are assigned with weights according to business
requirements, and the "goal" is assigned with a greater weight than that
assigned to the
"penalty kick" (the weights here are valuated between 0 and 1). In sports news
writing,
events are screened according to event weights, an event whose weight is
greater than the
preset threshold is preferentially selected to be written into the news title
section, so the
"goal" is written in the title section, whereas the "penalty kick" is written
in either the
abstract section or the text section.
[0168] Moreover, if an event further includes two sections as an abstract and
a text, and the
weight of the text event is higher than the weight of the abstract event, then
the "penalty
kick" will be preferentially written in the text section of the news.
[0169] In Embodiment 6, the sports news writing method based on a natural
language as
provided by the present application makes it possible to screen events,
whereby proper
events can be selected from many events according to requirement to provide
basis for
subsequent generation of news.
[0170] Embodiment 7
[0171] As shown in Fig. 7, in an embodiment of the present application,
screening the event
templates in step S105 includes the following steps:
[0172] Step S701 - obtaining the screened events and the codes corresponding
thereto, and the
codes to which all the event templates correspond in the events;
[0173] Step S702 - selecting one or more event template(s) with the maximum
number of slots
to serve as candidate event template(s); and
[0174] Step S703 - randomly selecting one event template from the candidate
event template(s)
to serve as the event template to be filled.
[0175] In steps S701-S703, the event template to be filled can be selected.
21
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
[0176] A specific example is taken below for explanation.
[0177] Table 3
[0178]
Project Number 1 2
Event 0010 1010
Event Template 0011 1011
[0179] As can be seen from Table 3, event 1 is coded 0010, event 2 is coded
1010, event template
1 is coded 0011, and event template 2 is coded 1011.
[0180] If event template 1 has one slot and event template 2 has two slots,
event template 2 with
the maximum number of slots is selected as the event template to be filled, so
event
template 2 serves as the unique one event template to be filled.
[0181] If event template 1 has one slot and event template 2 also has one
slot, since the event
template with the maximum number of slots is selected, so event template 1 and
event
template 2 each having one slot can each serve as the event template to be
filled, then one
event template is randomly selected from event template 1 and event template 2
to serve
as the event template to be filled, in other words, the event template to be
filled can be
either event template 1 or event template 2.
[0182] In Embodiment 7, the sports news writing method based on a natural
language as
provided by the present application makes it possible to screen event
templates, whereby
proper event templates can be selected from many event templates according to
requirement to provide basis for subsequent generation of news.
[0183] As shown in Fig. 8, in an embodiment of the present application, there
is further provided
22
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
a sports news writing device based on a natural language, to enhance the
diversity of
sentence patterns of articles and maximize information volume of articles, the
device
comprises:
[0184] an obtaining unit 801, for obtaining a corpus to be processed, an event
set, slots, and a
slot value to which each slot corresponds;
[0185] an event template marking unit 802, for marking event templates in the
corpus according
to each event in the event set, the slots, and the slot values;
[0186] a weight assigning unit 803, for performing weight assignment on each
event;
[0187] a coding unit 804, for coding each event, and types and the number of
the slots in the
event templates;
[0188] a screening unit 805, for screening the events and the event templates
according to the
weight of each event;
[0189] a news content generating unit 806, for matching and filling the
screened events and event
templates, and generating news content; and
[0190] a news content processing unit 807, for reprocessing the news content
to obtain the final
news content.
[0191] The device shown in Fig. 8 can correspondingly execute the contents in
the foregoing
method embodiment. Portions that are not described in detail in this
embodiment can be
inferred from the contents recited in the foregoing method embodiment, while
no
repetition is made in this context.
[0192] Refer below to Fig. 9 that is a view schematically illustrating a
structure suitable for
realizing the electronic equipment 90 according to an embodiment of the
current
disclosure. The electronic equipment in this embodiment of the current
disclosure can
include, but is not limited to include, such a mobile terminal as a mobile
phone, a
notebook computer, a digital broadcast receiver, a PDA (personal digital
assistant), a PAD
(portable Android device), a PIMP (portable multimedia player), an onboard
terminal
(such as an onboard navigation terminal), and such a fixed terminal as a
digital TV, a
23
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
desktop computer, etc. The electronic equipment shown in Fig. 3 is merely an
example,
and shall not be anyway restrictive of the function and range of use of the
embodiment
of the current disclosure.
[0193] As shown in Fig. 9, electronic equipment 90 can include a processing
device (such as a
central processor, a graphics processor, etc.) 901 capable of executing
various proper
actions and processes according to a program stored in a read-only memory
(ROM) 902
or a program loaded from a storage device 908 to a random-access memory (RAM)
903.
In RAM 903 are further stored various programs and data required for the
operation of
electronic equipment 90. Processing device 901, ROM 902 and RAM 903 are
connected
to one another via bus 904. Input/output (I/O) interface 905 is also connected
to bus 904.
[0194] Usually, the following systems can be connected to I/O interface 905:
including such an
input means 906 as a touch screen, a touch panel, a keyboard, a mouse, an
image sensor,
a microphone, an accelerometer, a gyroscope, etc.; such an output means 907 as
a liquid
crystal display (LCD), a loudspeaker, a vibrator, etc.; such a storge device
908 as a
magnetic disk, a hard disk, etc.; and a communication device 909.
Communication device
909 allows wireless or wired communication between electronic equipment 90 and
other
equipments to exchange data. Although an electronic equipment 90 possessing
various
devices is shown in the Figure, as should be understood, it is not required to
implement
or possess all the devices as shown. It is alternatively possible to implement
or possess
more or less devices.
[0195] Particularly, according to the embodiments of the current disclosure,
the processes
described above with reference to flowcharts can be realized as computer
software
programs. For instance, embodiments of the current disclosure include a
computer
program product that includes a computer program borne on a computer-readable
medium, and the computer program contains program codes for executing the
methods
shown in the flowcharts. In such an embodiment, the computer program can be
24
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
downloaded from the network and installed through communication device 909, or
installed from storage device 908, or installed from ROM 902. When the
computer
program is executed by processing device 901, it executes the aforementioned
functions
defined in the methods of the embodiments of the current disclosure.
[0196] The sports news writing method based on a natural language, and
corresponding device
and electronic equipment as provided by the present application make it
possible to
automatically generate news contents according to event templates extracted by
prior
analysis of great quantities of sports news and according to weight
assignments of users
themselves on events, whereby diversity of sentence patterns of articles is
enhanced and
information volume of articles is maximized on the one hand, highly
effectively
automatic writing of sports news articles is realized and the input of
manpower cost is
reduced on the other hand.
[0197] As should be noted, the computer-readable medium recited above in the
current
disclosure can be a computer-readable signal medium or a computer-readable
storage
medium or a random combination of the two. The computer-readable storage
medium
can for example be, but is not limited to be, an electric, magnetic, optical,
electromagnetic,
infrared, or semiconductor device, system or component, or any combination of
the above.
A more specific example of the computer-readable storage medium can include,
but is
not limited to include, an electrically connectible, portable computer
magnetic disk, a
hard disk, a random-access memory (RAM), a read-only memory (ROM), an erasable
programmable read-only memory (EPROM or flash memory), an optical fiber, a
portable
compact magnetic disk read-only memory (CD-ROM), an optical storage device, a
magnetic storage device having one or more conducting wire(s), or any suitable
combination of the above. In the current disclosure, the computer-readable
storage
medium can be any tangible medium containing or storing a program usable by or
in
combination with an instruction executing device, system, or component.
Moreover, in
the current disclosure, the computer-readable signal medium can include a data
signal
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
transmitted in a baseband or as part of a carrier wave, in which data signal
are borne
computer-readable program codes. The data signal thusly propagated can be
embodied in
plural forms, including, but not limited to, an electromagnetic signal, an
optical signal, or
any suitable combination thereof. The computer-readable signal medium can
further be
any other computer-readable medium than the computer-readable storage medium,
and
the computer-readable signal medium can transmit, propagate, or convey
programs for
use by or in combination with an instruction executing device, system, or
component.
The program codes contained in the computer-readable medium can be transmitted
via
any suitable medium, including, but not limited to, an electric wire, an
optical fiber, radio
frequency (RF) etc., or any suitable combination of the above.
[0198] The computer-readable medium can be either contained in the electronic
equipment, or
independent of, not installed in the electronic equipment.
[0199] The computer-readable medium carries therewith one or more program(s),
when the one
or more program(s) is/are executed by the electronic equipment, the electronic
equipment
is enabled: to acquire at least two internet protocol addresses; to transmit a
node
evaluation request including the at least two internet protocol addresses to a
node
evaluation equipment, which selects an internet protocol address from the at
least two
internet protocol addresses and returns the same; and to receive the internet
protocol
address returned by the node evaluation equipment, wherein the internet
protocol
addresses as acquired indicate boundary nodes in a content distribution
network.
[0200] Alternatively, the computer-readable medium carries therewith one or
more program(s),
when the one or more program(s) is/are executed by the electronic equipment,
the
electronic equipment is enabled: to receive a node evaluation request
including at least
two internet protocol addresses; to select an internet protocol address from
the at least
two internet protocol address; and to return the selected internet protocol
address, wherein
the internet protocol addresses as received indicate boundary nodes in a
content
26
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
distribution network.
[0201] One or more programming language(s) or a combination thereof can be
employed to write
the computer program codes for executing the operations of the current
disclosure, the
programming language(s) include(s) such an object-oriented programming
language as
Java, Smalltalk, C++, and further include(s) such a conventional procedural
programming
language as "C" language or a similar programming language. The program codes
can be
entirely executed on a user computer, partly executed on a user computer,
executed as an
independent software package, partly executed on a user computer and partly
executed
on a remote computer, or entirely executed on a remote computer or a server.
In the case
a remote computer is involved, the remote computer can be connected to the
user
computer via a randomly typed network, including a local area network (LAN) or
a wide
area network (WAN), or can be connected to an external computer (for example,
internet
connection can be supplied by an internet service provider).
[0202] The flowcharts and block diagrams in the accompanying drawings
illustrate systemic
frameworks, functions, and operations potentially realizable by the system,
method, and
computer program product according to the various embodiments of the current
disclosure. With respect thereto, each block in the flowcharts or the block
diagrams can
represent a module, a program segment, or a part of the codes, and the module,
the
program segment, or the part of the codes includes one or more executable
instruction(s)
for realizing designated logic function(s). As is also noticeable, in some
implementations
serving as substitutions, functions marked in the blocks can also occur in
sequences
different from those marked in the drawings. For instance, two blocks
expressed by
consecution can in practice be essentially parallelly executed, and they can
sometimes
also be executed according to inverse sequences, as these are decided by the
involved
functions. As should also be noted, each block in the block diagrams and/or
the flowcharts
and the combination of blocks in the block diagrams and/or the flowcharts can
be realized
by dedicated, hardware-based systems that execute designated functions or
operations,
27
Date Recue/Date Received 2022-06-22
CA 03165616 2022-06-22
and can be alternatively realized by a combination of dedicated hardware with
computer
instructions.
[0203] Units involved in the description of the embodiments of the current
disclosure can be
realized in the form of software, and can also be realized in the form of
hardware. The
name of a unit does not constitute any restriction to the unit itself under
certain
circumstances. For instance, a first obtaining unit can as well be referred to
as a "unit for
obtaining at least two intemet protocol addresses".
[0204] As should be understood, the various portions of the current disclosure
can be realized by
hardware, software, firmware, or a combination thereof.
[0205] What the above describes is merely directed to specific modes of
execution of the current
disclosure, but the protection scope of the current disclosure is not
restricted thereby. Any
change or replacement easily conceivable to persons skilled in the art within
the technical
range disclosed in the current disclosure shall all be covered by the
protection scope of
the current disclosure. Accordingly, the protection scope of the current
disclosure shall
be based on the protection scope of the attached Claims.
28
Date Recue/Date Received 2022-06-22