CN116910304B - Method, device, electronic equipment and storage medium for replacing video recommendation reason - Google Patents
Method, device, electronic equipment and storage medium for replacing video recommendation reason Download PDFInfo
- Publication number
- CN116910304B CN116910304B CN202311140182.8A CN202311140182A CN116910304B CN 116910304 B CN116910304 B CN 116910304B CN 202311140182 A CN202311140182 A CN 202311140182A CN 116910304 B CN116910304 B CN 116910304B
- Authority
- CN
- China
- Prior art keywords
- keyword
- video
- recall
- keywords
- click rate
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 61
- 238000012216 screening Methods 0.000 claims abstract description 16
- 238000004590 computer program Methods 0.000 claims description 21
- 238000012545 processing Methods 0.000 claims description 10
- 238000011084 recovery Methods 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 4
- 230000000694 effects Effects 0.000 abstract description 12
- 238000010586 diagram Methods 0.000 description 14
- 230000006870 function Effects 0.000 description 9
- 238000004891 communication Methods 0.000 description 6
- 241000219000 Populus Species 0.000 description 4
- 101100456831 Caenorhabditis elegans sams-5 gene Proteins 0.000 description 3
- 230000009471 action Effects 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 230000007774 longterm Effects 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 230000003190 augmentative effect Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/73—Querying
- G06F16/735—Filtering based on additional data, e.g. user or group profiles
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/25—Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
- H04N21/251—Learning process for intelligent management, e.g. learning user preferences for recommending movies
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
- H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
- H04N21/4667—Processing of monitored end-user data, e.g. trend analysis based on the log file of viewer selections
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/45—Management operations performed by the client for facilitating the reception of or the interaction with the content or administrating data related to the end-user or to the client device itself, e.g. learning user preferences for recommending movies, resolving scheduling conflicts
- H04N21/466—Learning process for intelligent management, e.g. learning user preferences for recommending movies
- H04N21/4668—Learning process for intelligent management, e.g. learning user preferences for recommending movies for recommending content, e.g. movies
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Theoretical Computer Science (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computing Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The application discloses a method and a device for recommending reasons of replacement video, electronic equipment and a storage medium. The method comprises the following steps: acquiring a recall source associated with the first video, wherein the recall source is used for screening the video of the specified type; determining candidate keywords associated with each recall source, wherein the candidate keywords comprise keywords with click rates greater than or equal to a preset threshold; combining the candidate keywords associated with each recall source to obtain a keyword combination meeting preset conditions, wherein the preset conditions are that the keyword combination comprises at least one entity word related to the first video content; calculating to obtain the score of each keyword combination according to the click rate of each candidate keyword in each keyword combination and the click rate of the corresponding recall source; and generating new recommendation reasons based on the keyword combination with the highest score, wherein the new recommendation reasons are used for replacing the current recommendation reasons of the first video. The embodiment of the application can improve the video recommendation effect and the video click rate.
Description
Technical Field
The present application relates to the field of video recommendation technologies, and in particular, to a method, an apparatus, an electronic device, and a computer readable storage medium for replacing a reason for video recommendation.
Background
In general, when a video platform recommends a video, a recommendation reason is often displayed for the video matching recommended, the recommendation reason is usually a short description of the main content or the bright spot of the video, the attention or resonance of a user is caused, and the proper recommendation reason can attract the attention of the user, so that the user is interested in the video and clicks to watch the video, and the video clicking rate can be improved.
Since the reason for recommending video is generally time-efficient, the initially set reason for recommending video may not match with video over time, or the current reason for recommending video may not attract the attention of the user, resulting in difficulty in maintaining the recommending effect. There are many ways to extract or generate the reason for video recommendation in the prior art, but there is no reasonable and efficient solution for how to maintain the recommendation effect. Such as: in Chinese patent literature with the name of ' a video recommended text generation method, a model training method and a related device ', the application number of which is 202210692931.7 ', text information corresponding to a target video is obtained, wherein the text information comprises description information and at least one tag information associated with the target video; and inputting the description information and at least one piece of label information into a pre-trained text generation model to obtain a recommended text of the target video. Although the recommended text can be automatically generated by using the training model, the timeliness and the recommended effect of the recommended text after a long time are not considered, and the problem of unsatisfactory recommended effect may occur.
Disclosure of Invention
In view of the above, embodiments of the present application provide a method, an apparatus, an electronic device, a computer-readable storage medium and a computer program product for replacing a reason for video recommendation, which are used for solving at least one technical problem.
The embodiment of the application provides a method for replacing video recommendation reasons, which is applied to a first video, wherein the recommendation reasons of the first video comprise one or more keywords; the method comprises the following steps: obtaining one or more recall sources associated with the first video, the recall sources being used to screen for specified types of video; determining one or more candidate keywords associated with each recall source, wherein the candidate keywords comprise keywords with click rates greater than or equal to a preset threshold; combining one or more candidate keywords associated with each recall source to obtain one or more keyword combinations meeting preset conditions, wherein the preset conditions are that the keyword combinations comprise at least one entity word related to the first video content; calculating to obtain the score of each keyword combination according to the click rate of each candidate keyword in each keyword combination and the click rate of the corresponding recall source, wherein the score of each keyword combination is used for representing the interested degree of a user on the corresponding keyword combination; and generating new recommendation reasons based on the keyword combination with the highest score, wherein the new recommendation reasons are used for replacing the current recommendation reasons of the first video.
According to the method of the embodiment of the application, the score of each keyword combination is determined according to the sum of products of the click rate of each candidate keyword in the keyword combination and the click rate of the corresponding recall source.
According to the method of the embodiment of the application, the score of each keyword combination is calculated by using the following formula:
wherein R (recallj) is a keyword combination score of the jth recall source of the first video including the ith candidate keyword; p (wi|recallj) is the click rate of the ith candidate keyword in the plurality of candidate keywords associated with the jth recall source; p (rectlj) is the click rate of the jth recall source; n is the number of candidate keywords associated with the jth recall source.
According to the method of the embodiment of the application, the preset threshold value comprises the click rate of the keyword type, and the click rate of the keyword type is the average value of the click rates of all keywords in the type of the keyword.
According to a method of an embodiment of the present application, before the determining one or more recall sources associated with the first video, the method further comprises: splitting the current recommendation reason of the first video into a plurality of keywords; if the click rate of at least one keyword in the plurality of keywords is greater than or equal to the keyword click rate threshold, the current recommendation reason is reserved, and replacement processing is not performed.
According to the method of the embodiment of the application, if at least one keyword in the plurality of keywords is the first-appearing keyword, the current recommendation reason is reserved, and no replacement processing is performed.
According to a method of an embodiment of the present application, before determining one or more candidate keywords associated with each recall source, the method further comprises: one or more keywords associated with each recall source are obtained, and if the keywords match the first video, the keywords are determined to be candidate keywords.
According to the method of the embodiment of the application, the types of the keywords comprise: time words, adjectives, and entity words.
In a second aspect, an embodiment of the present application provides an apparatus for recommending reason for replacing video, including: the acquisition module is used for acquiring one or more recall sources associated with the first video, wherein the recall sources are used for screening the video of the specified type; the determining module is used for determining one or more candidate keywords associated with each recall source, wherein the candidate keywords comprise keywords with click rate greater than or equal to a preset threshold value; the combination module is used for combining one or more candidate keywords associated with each recall source to obtain one or more keyword combinations meeting preset conditions, wherein the preset conditions are that the keyword combinations comprise at least one entity word related to the first video content; the calculation module is used for calculating and obtaining the score of each keyword combination according to the click rate of each candidate keyword in each keyword combination and the click rate of the corresponding recall source, wherein the score of each keyword combination is used for representing the interest degree of a user on the corresponding keyword combination; and the generation module is used for generating new recommendation reasons based on the keyword combination with the highest score and replacing the current recommendation reasons of the first video.
In a third aspect, an embodiment of the present application provides an electronic device, including: a processor and a memory storing computer program instructions; the electronic device, when executing the computer program instructions, implements the method as described above.
In a fourth aspect, embodiments of the present application provide a computer readable storage medium having stored thereon computer program instructions which, when executed by a processor, implement a method as described above.
By adopting the embodiment of the application, the score of the corresponding keyword combination is calculated according to the click rate of the candidate keywords and the click rate of the recall source, so that the interest degree of the user can be objectively and accurately reflected; based on the keyword combination with the highest score, new recommendation reasons are generated, the current recommendation reasons of the video are replaced, and the recommendation reasons which are not suitable any more can be replaced in time, so that the recommendation reasons of each time are the most popular or the most matched reasons at present, the recommended video can reach the best recommendation effect as far as possible, and the click rate is improved.
Drawings
In order to more clearly describe the technical solution of the embodiments of the present application, the following description briefly describes the drawings in the embodiments of the present application.
FIG. 1 is a schematic diagram of a system architecture according to an embodiment of the present application.
Fig. 2 is a flow chart of an alternate video recommendation reason method according to an embodiment of the present application.
Fig. 3 is a schematic diagram of a process for generating a reason for recommending video according to an embodiment of the present application.
Fig. 4 is a block diagram illustrating an alternative video recommendation reason device according to an embodiment of the present application.
Fig. 5 shows a schematic diagram of a hardware structure of an electronic device according to an embodiment of the present application.
Description of the embodiments
The principles and spirit of the present application will be described below with reference to several exemplary embodiments. It will be appreciated that such embodiments are provided to make the principles and spirit of the application clear and thorough, and enabling those skilled in the art to better understand and practice the principles and spirit of the application. The exemplary embodiments provided herein are merely some, but not all embodiments of the application. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the embodiments herein, are within the scope of the present application.
Embodiments of the present application relate to a terminal device and/or a server. Those skilled in the art will appreciate that embodiments of the application may be implemented as a system, apparatus, device, method, computer readable storage medium, or computer program product. Accordingly, the present disclosure may be embodied in at least one of the following forms: complete hardware, complete software, or a combination of hardware and software. According to embodiments of the present application, a method, apparatus, electronic device, computer-readable storage medium, and computer program product for replacing video recommendation reasons are claimed. Fig. 1 shows a schematic diagram of a system architecture according to an embodiment of the application. As shown in fig. 1, the system includes a terminal device 102 and a server 104. Wherein the terminal device 102 may comprise at least one of: smart phones, tablet computers, notebook computers, desktop computers, smart televisions, various wearable devices, augmented reality AR devices, virtual reality VR devices, and the like. The terminal device 102 may be provided with a client, for example, the client may be a client that specifically performs a specific function (such as an app), or a client embedded with multiple kinds of applets (different functions), or may be a client that logs in through a browser. The user may operate on the terminal device 102, for example, the user may open a client installed on the terminal device 102 and input an instruction through a client operation, or the user may open a browser installed on the terminal device 102 and input an instruction through a browser operation. After the terminal device 102 receives the instruction input by the user, request information including the instruction is transmitted to the server 104. The server 104 performs a corresponding process after receiving the request information, and then returns the process result information to the terminal device 102. User instructions are completed through a series of data processing and information interaction.
In this document, terms such as first, second, third, etc. are used solely to distinguish one entity (or action) from another entity (or action) without necessarily requiring or implying any order or relationship between such entities (or actions).
The following briefly describes related concepts and technical terms, etc. that may be involved in the embodiments of the present application.
The recall sources are used for screening out entries of specified types of videos from massive resources, and one video can be associated with a plurality of recall sources. One recall source is associated with one or more keywords. The reason for recommending a video is to describe the content or highlight of the video, attracting the user to click on the video. A video is associated with a recommendation reason. The recommended reasons comprise one or more keywords, and the types of the keywords comprise: time words, adjectives, and entity words.
The click rate of a video is the ratio of the number of times the video is clicked to the number of times it is exposed. Similarly, the click rate of the recall source is calculated in the same way as the click rate of the keyword. When the video is clicked by the number of times of +1, the number of clicks of the key words in the recall source and the recommendation reason associated with the video is also +1.
FIG. 2 shows a block flow diagram of an alternate video recommendation reason method according to an embodiment of the application, comprising the steps of:
s101: obtaining one or more recall sources associated with the first video, the recall sources being used to screen for specified types of video;
s102: determining one or more candidate keywords associated with each recall source, wherein the candidate keywords comprise keywords with click rates greater than or equal to a preset threshold;
s103: combining one or more candidate keywords associated with each recall source to obtain one or more keyword combinations meeting preset conditions, wherein the preset conditions are that the keyword combinations comprise at least one entity word related to the first video content;
s104: calculating to obtain the score of each keyword combination according to the click rate of each candidate keyword in each keyword combination and the click rate of the corresponding recall source, wherein the score of each keyword combination is used for representing the interested degree of a user on the corresponding keyword combination;
s105: and generating new recommendation reasons based on the keyword combination with the highest score, wherein the new recommendation reasons are used for replacing the current recommendation reasons of the first video.
The method is applied to the recommended video, and the method can calculate the current optimal recommended reason associated with the recommended video, so that the exposure effect of the recommended video is maximized, and the maximum click rate is achieved. Specifically, first, one or more recall sources associated with a first video are obtained, and one or more candidate keywords are screened out from each recall source. Because some keywords are time-efficient, they have previously met screening requirements and may not. Therefore, before generating new recommendation reasons each time, candidate keywords need to be screened again, so that timeliness of the candidate keywords is guaranteed, and a strong recommendation effect is achieved. Wherein the screening conditions may include: the keyword combination includes at least one entity word related to the first video content and the same type of keyword in the keyword combination appears only once.
The one or more candidate keyword combinations are then combined to obtain one or more keyword combinations, which may be a set of one or more keywords. By counting the exposure and click records of the first video in the last period of time, the click rate of the candidate key and the click rate of the recall source associated with the first video can be calculated. The click rate of the two can reflect the preference degree of the current user. Therefore, the score of the corresponding keyword combination is calculated according to the click rate of the candidate keywords and the click rate of the recall source, and the interested degree of the user can be objectively and accurately reflected. And finally, generating new recommendation reasons based on the keyword combination with the highest score, and replacing the current recommendation reasons of the first video. The new recommendation reason not only has the highest matching degree with the first video, but also can achieve the best recommendation effect and improve the click rate of the first video.
According to the embodiment of the application, optionally, the score of each keyword combination is determined according to the sum of products of the click rate of each candidate keyword in the keyword combination and the click rate of the corresponding recall source. Based on the principle that the higher the click rate score is, the higher the user interested degree is, the score of the keyword combination is obtained by the sum of products of the click rate of each candidate keyword and the click rate of the corresponding recall source, and the scores of the keyword combinations are compared, so that the keyword combination which is most interested by the user can be accurately and rapidly obtained, and the time for generating new recommendation reasons is shortened.
According to an embodiment of the present application, the score of each keyword combination is optionally calculated using the following formula：
Wherein R (recovery) j ) A keyword combination score for the jth recall source of the first video including the ith candidate keyword; p (w) i |recall j ) The click rate of the ith candidate keyword in the plurality of candidate keywords associated with the jth recall source; p (recovery) j ) The click rate of the jth recall source; n is the number of candidate keywords associated with the jth recall source. The magnitude of the score of the keyword combination can reflect the interest degree of the user, so that the recommendation effect of the keyword combination can be objectively and directly determined.
According to an embodiment of the present application, optionally, the preset threshold includes a keyword type click rate, where the keyword type click rate is an average value of click rates of all keywords in the type to which the keyword belongs. When the click rate of the keyword is lower than the click rate of the keyword type, the keyword is regarded as not welcome by users, and is a 'non-quality keyword'. Therefore, the candidate keyword is a "good keyword" after screening, and is a screening means necessary for generating a good recommendation reason.
Optionally, before the determining the one or more recall sources associated with the first video, the method further comprises: splitting the current recommendation reason of the first video into a plurality of keywords; if the click rate of at least one keyword in the plurality of keywords is greater than or equal to the keyword click rate threshold, the current recommendation reason is reserved, and replacement processing is not performed. When judging that the current recommendation reason is still popular with the majority of users, the replacement processing is not performed, and the recommendation time is shortened.
According to the embodiment of the application, optionally, if at least one keyword in the plurality of keywords is the first-appearing keyword, the current recommendation reason is reserved, and no replacement processing is performed. When the keyword appears for the first time, the click rate cannot be obtained because of no exposure record, the current recommendation reason is reserved, and after a period of time is recommended, whether replacement is needed is judged.
Optionally, before determining the one or more candidate keywords associated with each recall source, the method further comprises: one or more keywords associated with each recall source are obtained, and if the keywords match the first video, the keywords are determined to be candidate keywords. Matching, namely matching the keyword description with the content of the recommended video and the actual situation of the recommended video. And ensuring that the candidate keywords are matched with the first video, so that the generated recommendation reasons are consistent with the content or the actual situation of the first video, and improving the confidence.
Optionally, according to an embodiment of the present application, the types of the keywords include: time words, adjectives, and entity words. The keywords comprise various types of words, the content or the bright spots of the recommended video can be described from multiple dimensions, the attention of a user is attracted, and the exposure rate is improved.
The foregoing describes implementations and advantages of embodiments of the application in terms of a number of embodiments. The following describes in detail the specific processing procedure of the embodiment of the present application in conjunction with specific examples.
Fig. 3 is a schematic diagram illustrating a process of generating a reason for recommending video according to an embodiment of the present application. After receiving the user recommendation request, the server side recalls the videos interested by a plurality of users in the video library according to the user portrait as videos to be recommended, and then sends the videos to be recommended to the users. The video in the video to be recommended may be associated with a recommendation reason to attract the attention of the user. Before sending the recommended video with the recommended reason to the user, it is necessary to determine whether the current recommended reason needs replacement. For a certain video to be recommended, whether to replace the recommendation reason is judged according to the following three conditions.
1) And if the keywords in the recommendation reasons of the video to be recommended are in the candidate keyword example set, reserving the keywords, reserving the current recommendation reasons, and not replacing. The candidate keyword set is a keyword set in which the click rate of keywords is greater than a preset threshold (keyword type click rate).
2) If any keyword in the recommendation reasons is a new keyword, the current recommendation reason is reserved and no replacement is made. The new keyword appears for the first time in the existing keyword library.
3) If the reason for recommendation is lower than the corresponding preset threshold (click rate of keyword type), the reason for recommendation needs to be replaced. When all keywords in the recommendation reasons are lower than a preset threshold, the current recommendation reason is insufficient in attractive force, and the optimal recommendation effect cannot be achieved.
In connection with fig. 3, the following describes in detail the operation procedure of replacing the recommendation reason by using the embodiment of the present application, taking the video to be recommended as the dance video.
Firstly, acquiring the log information of the exposure click behavior of a user on the dance video, and extracting the recommended reason, recall source and interest point (if any) on each video according to the information collected during a certain period of time window (such as the last 10 data brushes). Referring to FIG. 2, the dance video is associated with recall sources R1-R5, and the recommended reasons for the dance video are T1-Tn. And splitting the recommended reasons T1-Tn into a plurality of keywords, and extracting the association relation between the recall sources R1-R5 and the split keywords.
There are several types of keywords, referring to fig. 2, numeral 1 is a time word, such as: yesterday, 7 days, long term …; number 2 is a term, for example: first name, first 90%, ten big … …; number 3 is an adjective, for example: is looking, cannot be seen, is looking … …; number 4 is entity word 1, for example: xx teacher, femto … …; number 5 is entity word 2, for example: podophylloic dance, cheongsam, fashion dance … …; the number n is the entity word n, dancing Sanjiang, bie Zhi Ji, ying Shang hong … ….
Then, the recall source click rate p (recovery), the keyword click rate p (w|recovery) and the keyword type click rate p (type (w)) are calculated according to the number of times the dance video is exposed and clicked in a certain time window period.
The recall source click rate p (recall) reflects the number of times that the recall source has been clicked within a period of time (e.g., the last 10 strokes in a session), and reflects the preference of a wide range of users. Each exposed video is associated with a different recall source, such as the dance video associated recall sources: mp 3-azalea, uid 1855210, subcat_265, etc., when the dance video is exposed 1 time, the associated recall sources are exposed 1 time, when the dance video is clicked 1 time, the recall sources contained in the dance video are clicked 1 time, and after all the associated videos of the recall sources are accumulated, the click rate p (recovery) of each recall source is calculated.
The clicking rate p (w|recycle), which is the ratio of the times that the corresponding keyword is clicked and exposed, can reflect the preference of the majority of users. For example, the keywords are: the keywords after splitting the reason documents are recommended by yesterday, the first name, the ghost dance, the dance Sanjiang and the like. As with the recall source click rate calculation method described above, when the dance video is clicked 1 time, the click times of all keywords in the corresponding recommended reason are +1, and the click rate p (w|recovery) of all individual keywords is counted in a period of time (in one session).
The keyword type click rate p (type (w)) is used to represent the average value of all keyword click rates in each type. For example, the keyword types corresponding to "yesterday", "7 days", "long term", which are abbreviated as keyword types, can be calculated as the average keyword click rate corresponding to the type to which the keyword belongs, that is, the keyword type click rate p (type (w)), by using the known click rate of the keyword.
And finally, carrying out secondary screening on the split keywords according to preset conditions to obtain candidate keywords, combining the candidate keywords into one or more keyword combinations, and selecting the optimal keyword combination from the one or more keyword combinations.
The first screening was performed using the following formula:
wherein p (w) i |recall j ) The click rate of the ith keyword in the plurality of keywords associated with the jth recall source is used for judging whether the jth recall source is the jth recall source; p (type (w) i )|recall j ) Click rate of the keyword type to which the i-th keyword belongs in the plurality of keywords associated with the j-th recall source. Using the formula [ (]1) High-quality keywords with higher average click rate than the type can be screened out.
The second screening was performed using the following formula:
wherein r (w) i I vid) indicates whether the i-th keyword is related to video vid, and when related, r (w) i I vid) =1, r (w when uncorrelated i I vid) =0. Using equation (2) can ensure that the new reason for recommendation is generated to match the dance video. For example, the i-th keyword is the first name, but the first video is not the first name, so that the keyword is inconsistent with the dance video, and the keyword needs to be removed, so that the authenticity of the recommendation reason is ensured.
After the secondary screening, combining the candidate keywords according to at least one entity word to obtain a plurality of keyword combinations, and calculating the score of each keyword combination according to a formula (3):
wherein R (recovery) j ) A keyword combination score of the ith candidate keyword is included in the jth recall source of the dance video; p (w) i |recall j ) The click rate of the ith candidate keyword in the plurality of candidate keywords associated with the jth recall source; p (recovery) j ) The click rate of the jth recall source; n is the number of candidate keywords associated with the jth recall source. The magnitude of the score of the keyword combination can reflect the interest degree of the user, so that the recommendation effect of the keyword combination can be objectively and directly determined.
The present application describes, by way of specific example, a process of calculating a keyword combination score using formula (3). For example, the title of the first video is: some dance video with playing at first sight is taken as an example of some 'hope you' original 32-step bouncing cycle jumping and playing. The recommended reasons are as follows: "dance music you want you" that many people are looking at.
The recommended reason document above is divided into two keyword types, namely (1) a "multi-person watching" of a collaborative viewing type and (2) a "hope you" of a dance entity type. And after searching the click rate of the 2 keyword types by the user, determining that the click rate of the keyword which is watched by multiple people is smaller than the click rate of the keyword type which is watched by the multiple people, and the click rate of the keyword which is watched by multiple people is smaller than the click rate of the keyword type which is watched by the dance entity, the original recommendation reason which is needed to be replaced is described.
All recall sources that acquire the video association by looking up are: mp3_ hope you, sub dance, tag 32 steps, follow _ poplar.
All keywords associated with "mp3 hope you" are acquired, and candidate keywords are obtained through secondary screening. The candidate keywords include: within "3 days", "first name". Because the recommendation reasons satisfy at least one entity, the recommendation reasons extracted for the "mp3 hope you" recall source are invalid.
And obtaining all keywords associated with 'subsubcat_dance', and obtaining candidate keywords through secondary screening. The candidate keywords include: within 3 days "," first name "," dance ". The click rates for "subtcat_dance" and multiple keywords are shown in table 1 below:
TABLE 1
Calculating the combination with the maximum keyword combination score according to the formula (3) as (bouncing dance, within three days, first name), wherein the recommended reason of the composition is as follows: the first dance is played within 3 days.
And obtaining all keywords associated in the step of tag_32, and obtaining candidate keywords through secondary screening. The candidate keywords include: "32 steps", "first name", "first 10 names". The click rates for "tag_32 steps" and multiple keywords are shown in table 2 below:
TABLE 2
The combination with the largest keyword combination score is calculated according to the formula (3) (32 steps, first 10) and the recommended reason of the composition is as follows: 32 steps into 10 before the door.
And obtaining all keywords associated with 'follow_poplar', and obtaining candidate keywords through secondary screening. The candidate keywords include: "Poplar someplace", "concerned". The click rate of "follow_popup" and a plurality of keywords is shown in table 3 below:
TABLE 3 Table 3
Calculating a combination with the maximum keyword combination score (some one is a poplar, concerned) according to a formula (3), wherein the recommendation reason of the composition is as follows: your work of a teacher who pays attention to a poplar.
From the above, the keyword combination is (dance, first name in three days) the highest score, and the recommendation reason is: "first dance in 3 days". The first dance in 3 days is used as a new recommendation reason to replace the current recommendation reason ' dance music hope you ' that a plurality of people are watching ', and the replacement work is completed.
Correspondingly, the present application also provides a device for recommending reason for replacing video, as shown in fig. 4, the device 100 for recommending reason for replacing video includes:
an acquisition module 110, configured to acquire one or more recall sources associated with the first video, where the recall sources are used to screen for a specified type of video;
a determining module 120, configured to determine one or more candidate keywords associated with each recall source, where the candidate keywords include keywords having a click rate greater than or equal to a preset threshold;
the combination module 130 is configured to combine one or more candidate keywords associated with each recall source to obtain one or more keyword combinations that conform to a preset condition, where the preset condition is that the keyword combinations include at least one entity word related to the first video content;
the calculation module 140 is configured to calculate a score of each keyword combination according to the click rate of each candidate keyword in each keyword combination and the click rate of the corresponding recall source, where the score of the keyword combination is used to represent the interest degree of the user on the corresponding keyword combination;
the generating module 150 is configured to generate a new recommendation reason based on the keyword combination with the highest score, and replace the current recommendation reason of the first video.
The electronic device in the embodiment of the application can be user terminal equipment, a server, other computing devices and a cloud server. Fig. 5 shows a schematic diagram of a hardware structure of an electronic device according to an embodiment of the present application, where the electronic device may include a processor 601 and a memory 602 storing computer program instructions, where the processor 601 implements the flow or functions of any of the methods of the embodiments described above when executing the computer program instructions.
In particular, the processor 601 may include a Central Processing Unit (CPU), or an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), or may be configured as one or more integrated circuits that implement embodiments of the present application. Memory 602 may include mass storage for data or instructions. For example, the memory 602 may be at least one of: hard Disk Drive (HDD), read-only memory (ROM), random-access memory (RAM), floppy Disk Drive, flash memory, optical Disk, magneto-optical Disk, magnetic tape, universal serial bus (Universal Serial Bus, USB) Drive, or other physical/tangible memory storage device. As another example, the memory 602 may include removable or non-removable (or fixed) media. For another example, memory 602 may be internal or external to the integrated gateway disaster recovery device. The memory 602 may be a non-volatile solid state memory. In other words, generally the memory 602 includes a tangible (non-transitory) computer-readable storage medium (e.g., a memory device) encoded with computer-executable instructions and when the software is executed (e.g., by one or more processors) may perform the operations described by the methods of embodiments of the application. The processor 601 implements the flow or functions of any of the methods of the above embodiments by reading and executing computer program instructions stored in the memory 602.
In one example, the electronic device shown in fig. 5 may also include a communication interface 603 and a bus 610. The processor 601, the memory 602, and the communication interface 603 are connected to each other through a bus 610 and perform communication with each other. The communication interface 603 is mainly used to implement communications between modules, apparatuses, units, and/or devices in the embodiments of the present application. Bus 610 includes hardware, software, or both, and may couple components of the online data flow billing device to each other. For example, the bus may include at least one of: accelerated Graphics Port (AGP) or other graphics bus, enhanced Industry Standard Architecture (EISA) bus, front Side Bus (FSB), hyperTransport (HT) interconnect, industry Standard Architecture (ISA) bus, infiniBand interconnect, low Pin Count (LPC) bus, memory bus, micro channel architecture (MCa) bus, peripheral Component Interconnect (PCI) bus, PCI-Express (PCI-X) bus, serial Advanced Technology Attachment (SATA) bus, video electronics standards Association local (VLB) bus, or other suitable bus. Bus 610 may include one or more buses. Although embodiments of the application describe or illustrate a particular bus, embodiments of the application contemplate any suitable bus or interconnection.
In connection with the methods of the above embodiments, embodiments of the present application also provide a computer-readable storage medium having stored thereon computer program instructions that, when executed by a processor, implement the flow or function of any of the methods of the above embodiments.
In addition, the embodiment of the present application further provides a computer program product, where the computer program product stores computer program instructions, and the computer program instructions implement the flow or the function of any one of the methods in the above embodiments when the computer program instructions are executed by a processor.
The foregoing exemplarily describes the flow diagrams and/or block diagrams of methods, apparatus, systems, and computer program products according to embodiments of the present application, and describes various aspects related thereto. It will be understood that each block of the flowchart illustrations and/or block diagrams, or combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions, special purpose hardware which perform the specified functions or acts, and combinations of special purpose hardware and computer instructions. For example, these computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the present application, enable the implementation of the functions/acts specified in the flowchart and/or block diagram block or blocks. Such a processor may be a general purpose processor, a special purpose processor, an application specific processor, or a field programmable logic circuit.
Functional blocks shown in the block diagrams of the embodiments of the present application can be implemented in hardware, software, firmware, or a combination thereof. When implemented in hardware, it may be, for example, an electronic circuit, an Application Specific Integrated Circuit (ASIC), suitable firmware, a plug-in, a function card, or the like; when implemented in software, are the programs or code segments used to perform the required tasks. The program or code segments can be stored in a memory or transmitted over transmission media or communication links through data signals carried in carrier waves. The code segments may be downloaded via computer networks such as the internet, intranets, etc.
It should be noted that the present application is not limited to the specific configurations and processes described above or shown in the drawings. The foregoing is merely specific embodiments of the present application, and it will be clearly understood by those skilled in the art that, for convenience and brevity of description, specific working processes of the described system, apparatus, module or unit may refer to corresponding processes in the method embodiments, and need not be repeated. It should be understood that the scope of the present application is not limited thereto, and any person skilled in the art may conceive various equivalent modifications or substitutions within the technical scope of the present application, which are intended to be included in the scope of the present application.
Claims (10)
1. A method for replacing a video recommendation reason, wherein the method is applied to a first video, the recommendation reason of the first video comprises one or more keywords, and the recommendation reason of the first video is used for describing video content or a highlight so that a user clicks the video; the method comprises the following steps:
obtaining one or more recall sources associated with the first video, the recall sources being used to screen for specified types of video;
determining one or more candidate keywords associated with each recall source, wherein the candidate keywords comprise keywords with click rates greater than or equal to a preset threshold;
combining one or more candidate keywords associated with each recall source to obtain one or more keyword combinations meeting preset conditions, wherein the preset conditions are that the keyword combinations comprise at least one entity word related to the first video content;
calculating to obtain the score of each keyword combination according to the click rate of each candidate keyword in each keyword combination and the click rate of the corresponding recall source, wherein the score of each keyword combination is used for representing the interested degree of a user on the corresponding keyword combination;
generating new recommendation reasons based on the keyword combination with the highest score, wherein the new recommendation reasons are used for replacing the current recommendation reasons of the first video;
and determining the score of each keyword combination according to the sum of products of the click rate of each candidate keyword in each keyword combination and the click rate of the corresponding recall source.
2. The method of claim 1, wherein the score for each keyword combination is calculated using the formula:
wherein R (recovery) j ) A keyword combination score for the jth recall source of the first video including the ith candidate keyword;
p(w i |recall j ) The click rate of the ith candidate keyword in the plurality of candidate keywords associated with the jth recall source;
p(recall j ) The click rate of the jth recall source; n is the number of candidate keywords associated with the jth recall source.
3. The method of claim 1, wherein the predetermined threshold comprises a keyword type click rate, the keyword type click rate being an average of click rates of all keywords in the type to which the keyword belongs.
4. The method of claim 1 or 3, wherein prior to the determining one or more recall sources associated with the first video, the method further comprises:
splitting the current recommendation reason of the first video into a plurality of keywords;
if the click rate of at least one keyword in the plurality of keywords is greater than or equal to the keyword click rate threshold, the current recommendation reason is reserved, and replacement processing is not performed.
5. The method of claim 4, wherein if at least one of the plurality of keywords is a first-occurring keyword, a current recommendation reason is reserved and no replacement process is performed.
6. The method of claim 1, wherein prior to determining the one or more candidate keywords associated with each recall source, the method further comprises:
one or more keywords associated with each recall source are obtained, and if the keywords match the first video, the keywords are determined to be candidate keywords.
7. The method of claim 1, wherein the type of keyword comprises: time words, adjectives, and entity words.
8. A replacement video recommendation reason device, comprising:
the acquisition module is used for acquiring one or more recall sources associated with the first video, wherein the recall sources are used for screening the video of the specified type;
the determining module is used for determining one or more candidate keywords associated with each recall source, wherein the candidate keywords comprise keywords with click rate greater than or equal to a preset threshold value;
the combination module is used for combining one or more candidate keywords associated with each recall source to obtain one or more keyword combinations meeting preset conditions, wherein the preset conditions are that the keyword combinations comprise at least one entity word related to the first video content;
the calculation module is used for calculating and obtaining the score of each keyword combination according to the click rate of each candidate keyword in each keyword combination and the click rate of the corresponding recall source, wherein the score of each keyword combination is used for representing the interest degree of a user on the corresponding keyword combination, and the score of each keyword combination is determined according to the sum of products of the click rate of each candidate keyword in each keyword combination and the click rate of the corresponding recall source;
and the generation module is used for generating new recommendation reasons based on the keyword combination with the highest score and replacing the current recommendation reasons of the first video, wherein the recommendation reasons of the first video are used for describing video contents or highlights so that a user clicks the video.
9. An electronic device, the electronic device comprising: a processor and a memory storing computer program instructions; the electronic device, when executing the computer program instructions, implements the method of any of claims 1-7.
10. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon computer program instructions which, when executed by a processor, implement the method according to any of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311140182.8A CN116910304B (en) | 2023-09-06 | 2023-09-06 | Method, device, electronic equipment and storage medium for replacing video recommendation reason |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311140182.8A CN116910304B (en) | 2023-09-06 | 2023-09-06 | Method, device, electronic equipment and storage medium for replacing video recommendation reason |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116910304A CN116910304A (en) | 2023-10-20 |
| CN116910304B true CN116910304B (en) | 2023-12-08 |
Family
ID=88356735
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311140182.8A Active CN116910304B (en) | 2023-09-06 | 2023-09-06 | Method, device, electronic equipment and storage medium for replacing video recommendation reason |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116910304B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104219575A (en) * | 2013-05-29 | 2014-12-17 | 酷盛(天津）科技有限公司 | Related video recommending method and system |
| CN116010630A (en) * | 2022-09-16 | 2023-04-25 | 北京奇艺世纪科技有限公司 | Real-time screening method and device for recommended video, electronic equipment and storage medium |
| CN116089567A (en) * | 2023-01-04 | 2023-05-09 | 浙江极氪智能科技有限公司 | Recommendation method, device, equipment and storage medium for search keywords |
| WO2023134407A1 (en) * | 2022-01-17 | 2023-07-20 | 北京字跳网络技术有限公司 | Content search method, apparatus and device, and medium |
| CN116521935A (en) * | 2022-01-20 | 2023-08-01 | 中国移动通信有限公司研究院 | Information recommendation method and device and electronic equipment |
-
2023
- 2023-09-06 CN CN202311140182.8A patent/CN116910304B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104219575A (en) * | 2013-05-29 | 2014-12-17 | 酷盛(天津）科技有限公司 | Related video recommending method and system |
| WO2023134407A1 (en) * | 2022-01-17 | 2023-07-20 | 北京字跳网络技术有限公司 | Content search method, apparatus and device, and medium |
| CN116521935A (en) * | 2022-01-20 | 2023-08-01 | 中国移动通信有限公司研究院 | Information recommendation method and device and electronic equipment |
| CN116010630A (en) * | 2022-09-16 | 2023-04-25 | 北京奇艺世纪科技有限公司 | Real-time screening method and device for recommended video, electronic equipment and storage medium |
| CN116089567A (en) * | 2023-01-04 | 2023-05-09 | 浙江极氪智能科技有限公司 | Recommendation method, device, equipment and storage medium for search keywords |
Non-Patent Citations (2)
| Title |
|---|
| Towards semantic and affective content-based video recommendation;Taiga Yoshida, et al;2013 IEEE International Conference on Multimedia and Expo Workshops (ICMEW);全文 * |
| 基于特征工程的视频点击率预测算法;匡俊;唐卫红;陈雷慧;陈辉;曾炜;董启民;高明;;华东师范大学学报(自然科学版)(第03期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116910304A (en) | 2023-10-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10567329B2 (en) | Methods and apparatus for inserting content into conversations in on-line and digital environments | |
| WO2020088058A1 (en) | Information generating method and device | |
| CN109509010B (en) | Multimedia information processing method, terminal and storage medium | |
| EP3117339A1 (en) | Systems and methods for keyword suggestion | |
| US10331685B2 (en) | Method and apparatus for sorting related searches | |
| WO2017206376A1 (en) | Searching method, searching device and non-volatile computer storage medium | |
| CN113383362B (en) | User identification method and related product | |
| CN111597446B (en) | Content pushing method and device based on artificial intelligence, server and storage medium | |
| CN112052297A (en) | Information generation method and device, electronic equipment and computer readable medium | |
| CN110209780B (en) | Question template generation method and device, server and storage medium | |
| CN104881447A (en) | Searching method and device | |
| CN108563713B (en) | Keyword rule generation method and device and electronic equipment | |
| CN113407775B (en) | Video searching method and device and electronic equipment | |
| CN109472028B (en) | Method and device for generating information | |
| CN116910304B (en) | Method, device, electronic equipment and storage medium for replacing video recommendation reason | |
| CN110971973A (en) | Video pushing method and device and electronic equipment | |
| CN114491093B (en) | Multimedia resource recommendation and object representation network generation method and device | |
| CN108363707B (en) | Method and device for generating webpage | |
| CN113365138B (en) | Content display method and device, electronic equipment and storage medium | |
| CN110659419B (en) | Method and related device for determining target user | |
| CN112672202B (en) | Bullet screen processing method, equipment and storage medium | |
| WO2017201904A1 (en) | Searching method, apparatus and device, and non-volatile computer storage medium | |
| CN109597873B (en) | Corpus data processing method and device, computer readable medium and electronic equipment | |
| CN113449212B (en) | Quality evaluation and optimization method, device and equipment for search results | |
| CN116578693B (en) | Text retrieval method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |