CN113626698A - Video recommendation method and device, electronic equipment and readable storage medium - Google Patents

Video recommendation method and device, electronic equipment and readable storage medium Download PDF

Info

Publication number
CN113626698A
CN113626698A CN202110900488.3A CN202110900488A CN113626698A CN 113626698 A CN113626698 A CN 113626698A CN 202110900488 A CN202110900488 A CN 202110900488A CN 113626698 A CN113626698 A CN 113626698A
Authority
CN
China
Prior art keywords
user
video
target
degree
preference degree
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.)
Pending
Application number
CN202110900488.3A
Other languages
Chinese (zh)
Inventor
王旭
王建兴
张晓明
张雪纯
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing QIYI Century Science and Technology Co Ltd
Original Assignee
Beijing QIYI Century Science and Technology Co Ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing QIYI Century Science and Technology Co Ltd filed Critical Beijing QIYI Century Science and Technology Co Ltd
Priority to CN202110900488.3A priority Critical patent/CN113626698A/en
Publication of CN113626698A publication Critical patent/CN113626698A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/22—Matching criteria, e.g. proximity measures

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Computation (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Artificial Intelligence (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computational Linguistics (AREA)
  • Multimedia (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)

Abstract

The embodiment of the invention provides a video recommendation method, a video recommendation device, electronic equipment and a readable storage medium, wherein the method comprises the following steps: acquiring play related parameters of a plurality of users aiming at the candidate video in a preset historical period; determining the target preference degree of each user to the viewed candidate video according to the playing related parameters; determining the target similarity of the first user and the second user according to the target preference degree of the first user to the first video and the target preference degree of the second user to the first video; determining the predicted preference degree of the first user to the second video according to the target similarity of the first user and the neighbor user, the first average target preference degree of the first user, the second average target preference degree of the neighbor user and the target preference degree of the neighbor user to the second video; determining a target video based on the predicted preference degree, and recommending the target video to the first user; the method and the device realize the pushing of the video resources really interested by the user for the user and also reduce the resources of the server.

Description

Video recommendation method and device, electronic equipment and readable storage medium
Technical Field
The present invention relates to the field of computer technologies, and in particular, to a video recommendation method and apparatus, an electronic device, and a readable storage medium.
Background
With the development of the internet, people's entertainment and learning ways no longer satisfy static resources such as characters, web pages, pictures, etc., and prefer to spend time on video websites to obtain required information and leisure and entertainment.
At present, video resources on a video website are various, and the classification is very different, and a user needs to search for video content interested by the user in a large amount, so that the problem caused by the fact that the user initiates a large amount of search requests and occupies too much resources of a server is solved. Therefore, how to push the video resources really interested by the user to the user, enrich and meet the personalized requirements of the user on the video resources, and reduce the resource occupation of the server becomes an urgent problem to be solved.
Disclosure of Invention
Embodiments of the present invention provide a video recommendation method, an apparatus, an electronic device, and a readable storage medium, so as to solve the problem how to push a video resource that is really interesting to a user for the user, enrich and meet personalized requirements of the user on the video resource, and reduce resource occupation of a server. The specific technical scheme is as follows:
in a first aspect of the present invention, there is provided a video recommendation method, including:
acquiring play related parameters of a plurality of users aiming at the candidate video in a preset historical period; the plurality of users includes a first user and a second user; the candidate videos include a first video and a second video; the first user is a user watching a first video but not watching a second video; the second user is a user watching the first video and the second video;
determining the target preference degree of each user to the viewed candidate video according to the playing related parameters; the target preference degree is in direct proportion to the playing related parameter;
determining the target similarity of the first user and the second user according to the target preference degree of the first user to the first video and the target preference degree of the second user to the first video; the target similarity is used for representing the similarity degree between the first user and the second user;
determining the predicted preference degree of the first user for the second video according to the target similarity of the first user and a neighbor user, the first average target preference degree of the first user, the second average target preference degree of the neighbor user and the target preference degree of the neighbor user for the second video; the neighbor user is a second user of which the target similarity with the first user is greater than a first preset threshold; the first average target like degree is an average value of the target like degrees of the first user to the first video; the second average target like degree of the neighbor user is an average value of the target like degrees of the neighbor user to the first video and the second video;
determining a target video in the second video based on the predicted like-degree, and recommending the target video to the first user. .
In a second aspect of the present invention, there is also provided a video recommendation apparatus, including:
the acquisition module is used for acquiring play related parameters of a plurality of users aiming at the candidate video in a preset history period; the plurality of users includes a first user and a second user; the candidate videos include a first video and a second video; the first user is a user watching a first video but not watching a second video; the second user is a user watching the first video and the second video;
a first determining module, configured to determine, according to the play-related parameter, a target preference degree of each user for the viewed candidate video; the target preference degree is in direct proportion to the playing related parameter;
a second determining module, configured to determine a target similarity between the first user and the second user according to a target like degree of the first user to the first video and a target like degree of the second user to the first video; the target similarity is used for representing the similarity degree between the first user and the second user;
a third determining module, configured to determine a predicted preference degree of the first user for the second video according to the target similarity between the first user and a neighboring user, the first average target preference degree of the first user, the second average target preference degree of the neighboring user, and the target preference degree of the neighboring user for the second video; the neighbor user is a second user of which the target similarity with the first user is greater than a first preset threshold; the first average target like degree is an average value of the target like degrees of the first user to the first video; the second average target like degree of the neighbor user is an average value of the target like degrees of the neighbor user to the first video and the second video;
and the recommending module is used for determining a target video in the second video based on the predicted preference degree and recommending the target video to the first user.
In another aspect of the present invention, there is also provided an electronic device, including a processor, a communication interface, a memory and a communication bus, where the processor, the communication interface, and the memory complete communication with each other through the communication bus;
a memory for storing a computer program;
and the processor is used for realizing the steps of the video recommendation method when executing the program stored in the memory.
In yet another aspect of the present invention, there is also provided a computer-readable storage medium on which a computer program is stored, which when executed by a processor, implements the video recommendation method described above.
According to the video recommendation method provided by the embodiment, play related parameters of a plurality of users aiming at candidate videos in a preset historical period are obtained; the plurality of users includes a first user and a second user; the candidate videos comprise a first video and a second video; the first user is a user watching the first video but not watching the second video; the second user is a user watching the first video and the second video; determining the target preference degree of each user to the viewed candidate video according to the playing related parameters; determining the target similarity of the first user and the second user according to the target preference degree of the first user to the first video and the target preference degree of the second user to the first video; determining the predicted preference degree of the first user to the second video according to the target similarity of the first user and the neighbor user, the first average target preference degree of the first user, the second average target preference degree of the neighbor user and the target preference degree of the neighbor user to the second video; and determining a target video in the second video based on the predicted preference degree, and recommending the target video to the first user. The target preference degree of the user to the candidate video is determined according to the playing related parameters of the user to the viewed candidate video, other users similar to the user are determined, and the preference degree of the user to the second video which is not viewed is predicted according to the target similarity degree and the target preference degree, so that the video resource which is really interested by the user is pushed to the user, and the personalized requirements of the user to the video resource are enriched and met. And because the user does not need to search interested video resources in the massive video resources, the resource occupation of the server is reduced.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below.
Fig. 1 is a flowchart illustrating steps of a video recommendation method according to an embodiment of the present invention;
fig. 2 is a schematic structural diagram of a video recommendation apparatus according to an embodiment of the present invention;
fig. 3 is a schematic structural diagram of an electronic device provided in an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be described below with reference to the drawings in the embodiments of the present invention.
Referring to fig. 1, fig. 1 is a flowchart illustrating steps of a video recommendation method according to an embodiment of the present invention. The method can be executed by a computer, a server and other electronic devices. The method comprises the following steps:
step 101, obtaining play related parameters of a plurality of users aiming at candidate videos in a preset history period; the plurality of users includes a first user and a second user; the candidate videos include a first video and a second video; the first user is a user watching a first video but not watching a second video; the second user is a user viewing the first video and the second video.
In this embodiment, the preset history period may be T days before the current time, where T is an integer greater than or equal to 1, for example, when T is equal to 1, the preset history period is one day before the current time, and when T is equal to 3, the preset history period is three days before the current time. A candidate video may refer to a union of videos viewed by multiple users. Exemplarily, if a first user watches videos a, b, and e, and a second user watches videos a, b, c, and d, the candidate videos are a, b, c, d, and e, where the first video is a video watched by both the first user and the second user, i.e., a and b; the second video is the video which is not viewed by the first user and is viewed by the second user, namely c and d. The playing related parameter may refer to a related parameter when the user watches the video, for example, the playing related parameter may be a number of clicks, a playing time length, a forwarding number, and the like of the candidate video by the user.
In this step, the electronic device may monitor the play related behavior of the user for the video resource in a manner such as embedding points, and further acquire a video viewed by the user in a preset history period as a candidate video, and simultaneously acquire play related parameters of the user for the candidate video, and subsequently determine a target preference degree of the user based on the play related parameters.
Step 102, determining the target preference degree of each user to the viewed candidate video according to the playing related parameters; the target like degree is proportional to the play-related parameter.
In the embodiment of the present invention, the target preference degree may refer to a preference degree of a candidate video viewed by a user, and the target preference degree may be represented in a form of a target preference index, where a higher playing related parameter of the user, for example, a higher number of clicks and a longer playing time, indicates that the user has a higher interest degree in the candidate video, and the target preference degree is higher. Specifically, when determining whether the user has viewed a certain candidate video, the determination may be made based on the click behavior of the user, and if the user has clicked the candidate video, it is determined that the user has viewed the candidate video.
Specifically, in this step, when determining the target preference degree, the specific calculation manner may be determined according to a specific category included in the play-related parameter. Illustratively, when the play-related parameters include the number of clicks and the play duration, the target preference of the user u for the candidate video c is Lu,cIs shown to be
Figure BDA0003199607240000051
The method comprises the steps that x represents the number of times that a user clicks a candidate video in a preset history period, y represents the playing time of the user for the candidate video in the preset history period, u represents an identifier of the user, c represents an identifier of the candidate video, the value range of c is an integer which is larger than or equal to 1 and smaller than or equal to P, and P is the total number of the candidate videos watched by the user u in the preset history period. For example, if the user 1 watches videos 1, 2, 4, 5, and 7 in a preset history period, the user 1 watches 5 candidate videos, the 5 first videos include videos 1, 2, 4, 5, and 7, and the target preference degree L of the user 1 for the videos 1 can be calculated according to the above calculation formula of the target preference degree1,1User 1's target preference level L for video 21,2User 1's target preference level L for video 41,4User 1's target like degree L for video 51,5User 1 target like degree L for video 71,7。
Similarly, if the user 2 watches the video 1, the video 2, the video 3, the video 5, the video 6 and the video 10 in the preset history period, the user 1 watches 6 candidate videos, and the 6 candidate videos include the video 1, the video 2, the video 3, the video 5, the video 6 and the video 10, and the preference index of the user 2 for each candidate video in the 6 candidate videos can be respectively calculated according to the target preference degree calculation formula.
103, determining a target similarity between the first user and the second user according to the target preference degree of the first user to the first video and the target preference degree of the second user to the first video; the target similarity is used for representing the similarity degree between the first user and the second user.
In this embodiment of the present invention, the target similarity may refer to a degree of similarity between the first user and the second user, and a higher degree of similarity indicates that the first user and the second user have a similar preference for the video. In this step, based on the first video viewed by both the first user and the second user, the electronic device may input the target preference degree of the first user for the first video, the target preference degree of the second user for the first video, the first average target preference degree of the first user, and the second average target preference degree of the second user into a preset similarity calculation formula, and determine the target similarity between the first user and the second user.
Specifically, when a target video needs to be recommended to a certain user among a plurality of users, the user is the first user. For example, when a target video needs to be recommended for user 1, user 1 is the first user. The second user includes at least one user of the plurality of users except the first user, and the second user needs to have intersection with the candidate video watched by the first user, so that the similarity between the first user and the second user can be determined based on the first video watched by the second user. By way of example, when the user 1 is a first user, users 2, 3, and 4 that intersect with the candidate video viewed by the first user may be second users.
For a clearer explanation of the present step, the target preference degrees of the user 1 on the candidate videos that have been watched by the user 1 and the target preference degrees of the user 2 on the candidate videos that have been watched by the user 2 are exemplarily described in the following table 1:
Figure BDA0003199607240000061
Figure BDA0003199607240000071
TABLE 1
It should be noted that the average target preference of the user
Figure BDA0003199607240000072
That is, the first average target preference degree of the first user and the second average target preference degree of the second user are equal to the ratio of the sum of the target preference degrees corresponding to the users to the total number of the target preference degrees corresponding to the users. From the data shown in Table 1 above, I corresponding to user 1 can be determined1,1、I1,2、I1,4、I1,5And I1,75 in total, then average target like-degree for user 1
Figure BDA0003199607240000073
Comprises the following steps: (0.03+0.04+0.01+0.05+0.02)/5, i.e., the average of the 5 target preference degrees corresponding to the user 1 is equal to 0.030. Also calculate the corresponding I of user 22,3、I2,1、I2,2、I2,5、I2,6And I2,10Total of 6 average target preference degrees
Figure BDA0003199607240000074
Is 0.042.
Similarly, the average target like degree corresponding to the user 3 is determined according to the target like degree of the candidate video viewed by the user 3
Figure BDA0003199607240000075
Based on the candidate video viewed by the user 4Target like degree, determining average target like degree corresponding to user 4
Figure BDA0003199607240000076
As shown in table 1 above, the first videos that are watched by both the user 1 and the user 2 in the preset history period include 2 videos, which are the video 1 and the video 2, and then the target similarity between the user 1 and the user 2 may be determined based on the 2 first videos and the average target preference degrees of the users 1 and 2.
Illustratively, how to determine the target similarity between two users is described below in conjunction with table 2 below, and in conjunction with the data in table 2 below, the target similarity between user 1 and user 2 may be determined:
Figure BDA0003199607240000077
TABLE 2
The target similarity between the first user and the second user is represented by s (u, v), then
Figure BDA0003199607240000078
Figure BDA0003199607240000079
The method comprises the steps of using u to represent a first user, using v to represent a second user, using I to represent an identifier of any one of first videos which are observed by the user u and the user v in a preset history period, and using I to represent any one of the first videos which are observed by the user u and the user v in the preset history periodu,iRepresenting a target degree of preference, I, of user u for a first video Iv,iRepresenting a target degree of preference, I, of a user v for a first video IuSet representing target like-degree corresponding to user u, IvSet (I) representing target preference degrees corresponding to user vu∩Iv) Representing a first video viewed by both user u and user v during a preset history period.
The first video watched by both user 1 and user 2 in the preset history period includes video 1 and video 2, and the target similarity of user 1 and user 2 is determined by combining the data shown in table 2 above
Figure BDA0003199607240000081
Figure BDA0003199607240000082
Similarly, the preference similarity s (1,3) of the user 1 and the user 3 is calculated according to the first video watched by the user 1 and the user 3 in the preset history period, and the target preference degree for the first video, the average target preference degree corresponding to the user 1, and the average target preference degree corresponding to the user 3. Similarly, the preference similarity s (1,4) of the user 1 and the user 4 can be calculated.
It should be noted that the specific calculation process of the target similarity may also be implemented in other manners, for example, using other calculation formulas, and the like.
104, determining the predicted preference degree of the first user for the second video according to the target similarity of the first user and a neighbor user, the first average target preference degree of the first user, the second average target preference degree of the neighbor user and the target preference degree of the neighbor user for the second video; the neighbor user is a second user of which the target similarity with the first user is greater than a first preset threshold; the first average target like degree is an average value of the target like degrees of the first user to the first video; the second average target like-degree of the neighboring user is an average of the target like-degrees of the neighboring users for the first video and the second video.
In the embodiment of the present invention, the neighbor user may refer to a second user whose target similarity with the first user is greater than a first preset threshold. The first preset threshold may be a preset similarity threshold, and when the first preset threshold is greater than the first preset threshold, the first user may be considered to be similar to the second user, and the preference of the first user to the video is closer, and the second user may be a neighbor user of the first user. The specific value of the first preset threshold may be determined based on actual requirements, and may be, for example, 0, that is, a second user having a target similarity greater than 0 with the first user may be a neighbor user of the first user.
Specifically, in this step, the preference degree of the neighbor user to the second video is approximately represented by the second average target preference degree of the neighbor user and the target preference degree of the neighbor user to the second video. Since there is a certain contingency in the target preference degrees of the neighboring users for each video, there may be a case where the target preference degree of the neighboring users for a certain second video is a high abnormal value, for example, it is assumed that the second average target preference degree of the neighboring users is 0.05, but for a certain second video, the target preference degree reaches 0.1, which obviously belongs to an abnormal value of the target preference degree of the neighboring users. At the moment, a parameter of the second average target preference degree of the neighbor user is introduced, and the preference degree of the neighbor user to the second video is combined with the target preference degree of the neighbor user to the second video to approximately represent the preference degree of the neighbor user to the second video, so that the interference of the target preference degree abnormal value of the neighbor user to the calculation result can be avoided, and the recommendation accuracy is improved.
And then, weighting the preference degree of the neighbor users to the second video by adopting the target similarity between the first user and the neighbor users, wherein the influence degree of each neighbor user to the first user can be determined based on the target similarity at the moment as the number of the neighbor users can be one or more. The higher the target similarity is, the larger the corresponding weight is, so that the neighbor users with the higher target similarity to the first user have a greater influence on the first user. After weighting the preference degrees of the neighboring users to the second video, a final target influence factor is obtained, and at this time, the target preference degree of the first user to the second video can be determined according to the target influence factor and the first average target preference degree of the first user.
In this step, after determining the target similarity between the first user and the second user, at least one neighbor user of the first user may be determined, and since the target similarity between the first user and the neighbor user is high, the preference of the first user and the neighbor user for the video is similar, the preference degree of the first user for the second video may be predicted based on the target preference degree of the neighbor user for the second video, and then the target video interested by the first user may be recommended to the first user based on the predicted preference degree.
And 105, determining a target video in the second video based on the predicted preference degree, and recommending the target video to the first user.
In this embodiment of the present invention, the target video may refer to a second video with a higher predicted preference degree of the user. After the predicted preference degree of the first user on the second video which is not watched is determined, the second video with the higher predicted preference degree of the first user can be used as the target video to be recommended to the user, and the recommendation accuracy is guaranteed.
In summary, in the video recommendation method provided by the embodiment of the present invention, play related parameters of a plurality of users for candidate videos in a preset history period are obtained; the plurality of users includes a first user and a second user; the candidate videos comprise a first video and a second video; the first user is a user watching the first video but not watching the second video; the second user is a user watching the first video and the second video; determining the target preference degree of each user to the viewed candidate video according to the playing related parameters; determining the target similarity of the first user and the second user according to the target preference degree of the first user to the first video and the target preference degree of the second user to the first video; determining the predicted preference degree of the first user to the second video according to the target similarity of the first user and the neighbor user, the first average target preference degree of the first user, the second average target preference degree of the neighbor user and the target preference degree of the neighbor user to the second video; and determining a target video in the second video based on the predicted preference degree, and recommending the target video to the first user. The target preference degree of the user to the candidate video is determined according to the playing related parameters of the user to the viewed candidate video, other users similar to the user are determined, and the preference degree of the user to the second video which is not viewed is predicted according to the target similarity degree and the target preference degree, so that the video resource which is really interested by the user is rapidly pushed to the user, and the personalized requirements of the user to the video resource are enriched and met. In addition, the user does not need to search interested video resources in the massive video resources, so that the resource occupation of the server is reduced.
Optionally, in this embodiment of the present invention, step 101 may specifically include the following steps 1011 to 1013:
step 1011, obtaining the play-related behavior logs of the users, and writing the play-related behavior logs into a preset database.
In the embodiment of the present invention, playing the related behavior log may refer to various operation behaviors of the user on the video on the application program or the webpage, such as clicking, playing, praise, comment, forwarding, and the like. The preset database may be a preset database for storing logs, and specifically may be a data warehouse tool Hive or the like.
Specifically, in this step, the electronic device may perform point burying on the play-related behavior of the user, such as a video play behavior and a video click behavior, Extract, convert, and Load the log based on a data warehouse technology (ETL) by using a preset processing program (e.g., a Flink program), and finally write the sorted play-related behavior log data into a preset database, thereby completing the acquisition and the obtaining of the log data.
Step 1012, screening out play-related behaviors of the plurality of users for the candidate video from the play-related behavior log according to a preset history sub-period; the preset history period is composed of a plurality of continuous preset history sub-periods.
In this embodiment of the present invention, the preset history sub-period may refer to each sub-period within the preset history period, for example, when the preset history period is one day, the preset history sub-period may be one hour. The play-related behavior may refer to a click behavior, a play behavior, a forward behavior, and the like of the user with respect to the candidate video.
In this step, after the related behavior logs of the multiple users are obtained, because the number of the users is large and the types of data in each obtained log are also large, the calculation amount for directly processing is too large, and more system resources are occupied, the data needs to be screened and sorted. Specifically, the electronic device may first filter and count the play-related parameters in the play-related behavior log by using the data in the preset history sub-period as a batch, so as to obtain the play-related parameters in the preset history sub-period. For example, when the preset history period is one day, the data amount in 24 hours per day is directly counted, and the playing related parameters in a preset history sub-period, for example, in 1 hour, may be filtered and counted first, and then the data in 24 hours is aggregated and counted.
Illustratively, table 3 shows a detailed data type of a play-related behavior log according to an embodiment of the present invention.
Figure BDA0003199607240000111
Figure BDA0003199607240000121
Figure BDA0003199607240000131
TABLE 3
As can be seen from table 3, the types of detailed data in the play-related behavior log generated in real time are many, the data is complex, the amount of data to be directly processed is too large, and the detailed data needs to be screened when the play-related parameters are screened and obtained, so as to reduce the amount of data calculation.
Illustratively, table 4 shows data types of the play-related parameters after the play-related behavior log is filtered according to the embodiment of the present invention.
Figure BDA0003199607240000132
Figure BDA0003199607240000141
TABLE 4
As can be seen from table 4, after the play-related behavior log is screened, only the data related to the play-related parameters may be screened, so that the data is simplified, and the amount of calculation of the data is reduced.
And 1013, performing aggregation statistics on the playing related parameters in the preset history sub-periods according to a preset history period to obtain the playing related parameters in the preset history period.
In the embodiment of the invention, after the playing related parameters in each preset history sub-period are acquired, the user can perform aggregation statistics. Specifically, the playing related parameters in a plurality of consecutive preset history sub-periods may be aggregated to obtain the playing related parameters in the whole preset history period. For example, when the preset history sub-period is 1 hour, and the preset history period is one day, the electronic device may first obtain, based on the play-related behavior log, play-related parameters of the user for the candidate video within one hour, and then perform aggregation statistics on the play-related parameters for 24 hours a day, so as to obtain the play-related parameters of the multiple users for the candidate video within the preset history period, that is, within 1 day.
In the embodiment of the invention, play-related behavior logs of a plurality of users are obtained and are written into a preset database; screening out play related parameters of a plurality of users aiming at the candidate videos from play related behavior logs according to a preset historical sub-period; the preset history period consists of a plurality of continuous preset history sub-periods; and carrying out aggregation statistics on the playing related parameters in the preset history sub-periods according to the preset history period to obtain the playing related parameters in the preset history period. Therefore, by screening and carrying out sectional statistics on the logs, the data calculation amount is reduced, the calculation efficiency is improved, the time consumed by recommending videos for users is further shortened, and quick recommendation is facilitated.
Optionally, in this embodiment of the present invention, before step 1012, the video recommendation method may further include the following steps S11 to S12:
step S1, determining a target hot video and a video identifier of the target hot video; the target hot video is a video with the average playing time length in at least one preset history period being greater than a second preset threshold value.
In the embodiment of the present invention, the target hot video may refer to a video with a large playing amount in a recent period of time. The recent period of time may refer to one preset history period, or may refer to a plurality of preset history periods, for example, it may be a day, a week, or a month, and this is not specifically limited in the embodiment of the present invention. The second preset threshold may refer to a preset critical threshold of the playing time. If the video playing amount is larger than the first preset threshold, it can be determined that the video playing amount is larger and the video is a target hot video. Specifically, when determining whether the video is the target thermal video, statistics may be performed only on the data of the playing time length in the log, the average playing time length of the video in at least one preset history period is determined, and the video may be determined to be the target thermal video when it is determined that the average playing time length is greater than a second preset threshold.
Generally speaking, the overall playing related parameters of the target hot video are obviously higher than those of other videos, and are often displayed on a remarkable position of a video platform, so that a user can visually acquire the target hot video, and the target hot video does not need to be recommended to the user. Moreover, because the playing amount of the target hot video is large, a large amount of play-related behavior log data can be generated, and the calculation amount in the subsequent screening statistics is also large. In this step, the average playing time of the video is determined according to the specific data of the playing time, the target thermal video is further determined through comparison with a second preset threshold, and the target thermal video can be deleted from the candidate video subsequently, so that a large amount of log data generated by the target thermal video is prevented from being processed, and the execution efficiency of the algorithm is further improved.
Of course, in this step, the target thermal video may also be determined in other manners, for example, the target thermal video is determined based on the number of clicks, a video with the number of clicks in the last 30 days being ranked in the top 15 digits is used as the target thermal video, and the like, and the flexible setting may be specifically performed based on actual needs, which is not limited in this embodiment of the disclosure.
Step S2, based on the video identifier of the target thermal video, filtering out the play-related behavior log related to the target thermal video from the play-related behavior log.
In the embodiment of the invention, after the target thermal video is determined, the target thermal video can be deleted from the candidate video, specifically, the play-related behavior log related to the target thermal video can be removed from the play-related behavior log stored in the preset database, so that the log data related to the target thermal video does not need to be processed subsequently, and the data processing amount is reduced.
In the embodiment of the invention, a target thermal video and a video identifier of the target thermal video are determined; the target hot video is a video with the average playing time length in at least one preset history period being greater than a second preset threshold value; and filtering the play-related behavior log related to the target thermal video from the play-related behavior log based on the video identification of the target thermal video. Therefore, by filtering the relevant data of the target hot video, the data processing amount can be reduced, the execution efficiency of the algorithm is improved, and the quick recommendation of the user is realized.
Optionally, in this embodiment of the present invention, before step 103, the video recommendation method may further include the following steps S21 to S22:
step S21, determining an average value of target like degrees of all users who have viewed the candidate video for the candidate video, to obtain an average like degree.
In the embodiment of the present disclosure, the average degree of preference may refer to an average value of target degrees of preference of a user who has viewed a particular candidate video for the candidate video.
Specifically, in this step, when calculating the average preference degree, the total preference degree of the user for the candidate video may be determined based on the target preference degree of the user who has watched the candidate video for the candidate video, and then the total preference degree may be obtained by dividing by the number of the users who have watched the candidate video. Illustratively, the total preference index of multiple users for a candidate video is represented by SLcIs shown to be
Figure BDA0003199607240000161
The example is given by taking an example that a plurality of users include 4 users, the 4 users include user 1, user 2, user 3 and user 4, and the preference index of user 1 to candidate video 1 is L1,1The preference index of the user 2 to the candidate video 1 is L2,1The preference index of the user 3 to the candidate video 1 is L3,1User 4 preference index L for candidate video 14,1The total preference index of the 4 users for the candidate video 1 is represented as SL1，SL1＝L1,1+L2,1+L3,1+L4,1. The overall preference index for video 2 for the 4 users can also be calculated and expressed as SL2And so on. The total preference indication can then be divided by the number of users, i.e. SL1And/4, obtaining the average preference degree.
It should be noted that the overall preference index SL for video 1 of the 4 users is calculated1Thereafter, if the user's target like degree for video 1 uses Iu,cIs shown to be
Figure BDA0003199607240000162
I.e. by means of a formula
Figure BDA0003199607240000163
To Lu,cAnd (6) carrying out normalization processing. For example,
Figure BDA0003199607240000164
therefore, the data are processed in a unified way, the subsequent calculation can be facilitated, and the data calculation is more convenient and quickerIt is quick.
And step S22, determining the candidate video as the target cold video and deleting the target cold video from the candidate video when the average degree of preference is lower than a third preset threshold.
In this embodiment of the present invention, the third preset threshold may be a preset threshold of average preference degree, and when the third preset threshold is lower than the third preset threshold, it may be determined that the candidate video is the target cold video. The target cold video may be a video with a low overall user preference degree, and since the preference degrees of the plurality of users who have viewed the candidate video to the candidate video are low, the probability that the current first user is not interested in the video is high, and the current user may not be recommended.
In the step, after the target cold video is determined, the target cold video is deleted from the candidate videos, and then when recommendation is performed for the user, the target cold video is not recommended any more so as to avoid recommending uninteresting video resources for the user.
In the embodiment of the invention, the average value of the target preference degrees of all users watching the candidate videos for the candidate videos is determined for each candidate video, and the average preference degree is obtained; and in the case that the average preference degree is lower than a third preset threshold value, determining the candidate video as the target cold video and deleting the target cold video from the candidate video. Therefore, target cold videos which are not interested by the user as a whole are filtered from the candidate videos, the calculation amount of the subsequent preference prediction degree can be reduced, the range of recommended videos can be narrowed, and the video recommendation accuracy is further improved.
Optionally, in this embodiment of the present invention, step 104 may be implemented by the following step S31:
step S31, determining a predicted preference degree of the first user for the second video according to the first average target preference degree and the target influence factor of the first user; wherein the target impact factor is determined based on the target similarity between the first user and the neighboring user, and the target like degree of the neighboring user to the second video and the second average target like degree of the neighboring user are weighted.
In the embodiment of the present invention, the target influence factor may refer to an adjustment value during calculation of the preset preference degree, and is a numerical value obtained by weighting the preference degree of the neighboring user on the second video, so that the overall preference degree of the neighboring user on the second video can be reflected. Because the target similarity between the first user and the neighbor user is high, the second video which the neighbor user likes integrally can be considered, and the probability that the first user likes the second video is high.
Specifically, in this step, the preference degree of the neighbor user for the second video may be approximately determined based on the target preference degree of the neighbor user for the second video and the second average target preference degree of the neighbor user, and meanwhile, the interference of the abnormal value may be eliminated. And then determining a weighting weight according to the target similarity of the first user and the neighbor user, wherein the higher the target similarity is, the more similar the interest and the preference of the first user and the neighbor user to the video are, the larger the weighting weight is. And then, according to the determined weighting weight, weighting the preference degree of at least one neighbor user to the second video to obtain a target influence factor, wherein the target influence factor can represent the overall preference degree of the neighbor user to the second video, and the specific calculation mode can be a calculation mode such as weighted summation. Finally, the predicted preference level of the first user for the second video may be determined based on the target impact factor and the first average target preference level of the first user. By introducing the parameter of the first average target like degree, the interference of the abnormal value of the target like degree of the first user can be eliminated and the calculation accuracy can be improved by calculating the average value of the target like degree of the video viewed by the first user. .
Optionally, in this embodiment of the present invention, step 104 may be specifically implemented by steps 1041 to 1042 as follows:
step 1041, based on the target similarity, screening a second user whose target similarity with the first user is greater than a first preset threshold as the neighbor user, and adding the neighbor user to a neighbor user set N.
In this embodiment of the present invention, the neighbor user set N may refer to a set of second users whose similarity to the first user is greater than a first preset threshold. By determining the neighbor users, the preference degree of the first user for the second video can be predicted subsequently based on the preference degree of at least one neighbor user for the second video.
Step 1042, for each second video, inputting the target similarity between the first user and a neighboring user, the first average target like degree, the second average target like degree of the neighboring user, and the target like degree of the neighboring user to the second video into a target prediction function, and determining the predicted like degree of the first user to the second video;
wherein the target prediction function is:
Figure BDA0003199607240000181
wherein p isu,iIn order to predict the preference degree, i is a second video to be predicted; u is a first user, and N is a neighbor user in the neighbor user set N; s (u, n) is the target similarity of the first user u and the neighbor user n;
Figure BDA0003199607240000182
the first average target preference level;
Figure BDA0003199607240000183
the second average target preference level; r isn,iA target like-degree of the second video i to be predicted for the neighboring user n.
In an embodiment of the present invention, the target prediction function may be used to determine a predicted preference index of the first user for the second video. Specifically, step 1042 is described below with reference to the above example:
if the candidate videos watched by the multiple users in the preset history period include 10 videos including video 1, video 2, video 3, video 4, video 5, video 6, video 7, video 8, video 9 and video 10. Still taking the above-mentioned examples that the plurality of users include user 1, user 2, user 3, and user 4 together, then the second video that is not viewed by user 1 includes video 3, video 6, video 8, video 9, and video 10, and the predicted preference level of user 1 for video 3, video 6, video 8, video 9, and video 10 can be determined.
For example, taking the first user as user 1 as an example, the data of the target similarity between user 1 and all the second users is shown in the following table 5:
Figure BDA0003199607240000184
Figure BDA0003199607240000191
TABLE 5
If the first preset threshold is equal to 0, since the target similarity between the user 1 and the user 3 is equal to 0.000, and the target similarity is equal to the first preset threshold, the user 3 needs to be filtered; if the target similarity between the user 1 and the user 4 is equal to-8.012 and smaller than the first preset threshold, the user 4 needs to be filtered out, that is, only the user 2 is the neighbor user whose target similarity with the first user is larger than the first preset threshold. In this case, the neighbor user set N includes only user 2, and accordingly, is based on the target prediction function, i.e.
Figure BDA0003199607240000192
The predicted preference degree of the user 1 for the video 6 can be known by substituting the corresponding data
Figure BDA0003199607240000193
Wherein, I2,6Representing the target preference level of the user 2 for the video 6, with reference to the data, I, shown in Table 1 above2,6Equal to 0.01, and as described in connection with the above example, the average target like-degree of the video viewed by the user 2 during the predetermined history period
Figure BDA0003199607240000194
Equal to 0.042.
In the embodiment of the invention, based on the target similarity, screening a second user with the target similarity larger than a first preset threshold with the first user as a neighbor user, and adding the neighbor user to a neighbor user set N; and for each second video, inputting the target similarity of the first user and the neighbor user, the first average target preference degree, the second average target preference degree of the neighbor user and the target preference degree of the neighbor user to the second video into a target prediction function, and determining the predicted preference degree of the first user to the second video. Therefore, based on the target similarity and the target preference degree of the neighbor user to the second video, the preference degree of the first user to the second video which is not watched can be determined quickly, and the video recommendation accuracy is improved.
Optionally, in the embodiment of the present invention, step 105 may be specifically implemented by the following steps 1051 to 1052:
and 1051, sorting the second video in a descending order according to the predicted preference degree.
In the embodiment of the present invention, after the predicted preference degree of the first user for the second video is determined, the second video may be sorted in a descending order according to the order from high to low of the predicted preference degree based on the predicted preference degree.
Step 1052, selecting a preset number of second videos sequenced at the top as the target videos according to the sequencing sequence.
In this embodiment of the present invention, the preset number may refer to a preset number, for example, the preset number may be 5, 10, 15, and the like, and may be specifically determined according to actual requirements, which is not limited in this embodiment of the present invention. After the second videos are sorted according to the predicted preference degree, the first preset number of second videos can be selected as target videos and recommended to the user, and therefore the user interest of the recommended videos can be guaranteed to be high.
Illustratively, still taking the user 1 as the first user as an example, the pair of videos 3, 6 and view of the user 1 is determinedAfter the predicted preference degrees of the video 8, the video 9 and the video 10 are predicted, the predicted preference degrees can be sorted in a descending order, a preset number of predicted preference degrees are sequentially selected from the largest predicted preference degree, and the second video corresponding to the selected predicted preference degree is used as the target video. For example, if the predetermined number is equal to 3, p1,6>p1,3>p1,8>p1,9>p1,10Then the second video 6, the video 3, and the video 8 are taken as target videos, and the video 6, the video 3, and the video 8 are recommended to the user 1.
In the embodiment of the invention, according to the predicted preference degree, the second videos are sorted in a descending order according to the predicted preference degree; and selecting a preset number of second videos sequenced at the top as target videos according to the sequencing sequence. Therefore, the accuracy of recommending videos for the user can be ensured by screening the second video with the predicted preference degree ranked in the front as the target video, and the video resources really interested by the user can be pushed for the user.
Referring to fig. 2, fig. 2 is a schematic structural diagram of a video recommendation apparatus 200 according to an embodiment of the present invention, the apparatus 200 is disposed in an electronic device such as a computer, and the apparatus 200 includes:
an obtaining module 201, configured to obtain play related parameters of multiple users for a candidate video in a preset history period; the plurality of users includes a first user and a second user; the candidate videos include a first video and a second video; the first user is a user watching a first video but not watching a second video; the second user is a user watching the first video and the second video;
a first determining module 202, configured to determine, according to the play-related parameter, a target preference degree of each user for the viewed candidate video; the target preference degree is in direct proportion to the playing related parameter;
a second determining module 203, configured to determine a target similarity between the first user and the second user according to the target preference degree of the first user for the first video and the target preference degree of the second user for the first video; the target similarity is used for representing the similarity degree between the first user and the second user;
a third determining module 204, configured to determine a predicted preference degree of the first user for the second video according to the target similarity between the first user and a neighboring user, the first average target preference degree of the first user, the second average target preference degree of the neighboring user, and the target preference degree of the neighboring user for the second video; the neighbor user is a second user of which the target similarity with the first user is greater than a first preset threshold; the first average target like degree is an average value of the target like degrees of the first user to the first video; the second average target like degree of the neighbor user is an average value of the target like degrees of the neighbor user to the first video and the second video;
a recommending module 205, configured to determine a target video in the second video based on the predicted like-degree, and recommend the target video to the first user.
In summary, the video recommendation apparatus provided in the embodiment of the present invention obtains the play related parameters of the candidate videos for the plurality of users in the preset history period; the plurality of users includes a first user and a second user; the candidate videos comprise a first video and a second video; the first user is a user watching the first video but not watching the second video; the second user is a user watching the first video and the second video; determining the target preference degree of each user to the viewed candidate video according to the playing related parameters; determining the target similarity of the first user and the second user according to the target preference degree of the first user to the first video and the target preference degree of the second user to the first video; determining the predicted preference degree of the first user to the second video according to the target similarity of the first user and the neighbor user, the first average target preference degree of the first user, the second average target preference degree of the neighbor user and the target preference degree of the neighbor user to the second video; and determining a target video in the second video based on the predicted preference degree, and recommending the target video to the first user. The target preference degree of the user to the candidate video is determined according to the playing related parameters of the user to the viewed candidate video, other users similar to the user are determined, and the preference degree of the user to the second video which is not viewed is predicted according to the target similarity degree and the target preference degree, so that the video resource which is really interested by the user is pushed to the user, and the personalized requirements of the user to the video resource are enriched and met. And because the user does not need to search interested video resources in the massive video resources, the resource occupation of the server is reduced.
Optionally, the obtaining module 201 is specifically configured to:
acquiring play-related behavior logs of the users, and writing the play-related behavior logs into a preset database;
screening out playing related parameters of the plurality of users aiming at the candidate videos from the playing related behavior log according to a preset historical sub-period; the preset history period consists of a plurality of continuous preset history sub-periods;
and carrying out aggregation statistics on the playing related parameters in the preset history sub-periods according to the preset history period to obtain the playing related parameters in the preset history period.
Optionally, the obtaining module 201 is further specifically configured to:
determining a target thermal video and a video identifier of the target thermal video; the target hot video is a video with the average playing time length in at least one preset history period being greater than a second preset threshold value;
and filtering the play-related behavior log related to the target hot video from the play-related behavior log based on the video identification of the target hot video.
Optionally, the apparatus 200 further includes:
the fourth determining module is used for determining the average value of the target preference degrees of all the users who watch the candidate videos for the candidate videos to obtain the average preference degree;
and the deleting module is used for determining the candidate video as the target cold video and deleting the target cold video from the candidate video under the condition that the average like degree is lower than a third preset threshold.
Optionally, the third determining module 204 is specifically configured to:
screening a second user with the target similarity larger than a first preset threshold value with the first user as the neighbor user based on the target similarity, and adding the neighbor user to a neighbor user set N;
for each second video, inputting the target similarity of the first user and a neighbor user, the first average target like degree, the second average target like degree of the neighbor user and the target like degree of the neighbor user to the second video into a target prediction function, and determining the predicted like degree of the first user to the second video;
wherein the target prediction function is:
Figure BDA0003199607240000231
wherein p isu,iIn order to predict the preference degree, i is a second video to be predicted; u is a first user, and N is a neighbor user in the neighbor user set N; s (u, n) is the target similarity of the first user u and the neighbor user n;
Figure BDA0003199607240000232
the first average target preference level;
Figure BDA0003199607240000233
the second average target preference level; r isn,iA target like-degree of the second video i to be predicted for the neighboring user n.
Optionally, the recommending module 205 is specifically configured to:
according to the predicted preference degree, sorting the second video in a descending order according to the predicted preference degree;
and selecting a preset number of second videos sequenced in the front as the target videos according to the sequencing sequence.
An embodiment of the present invention further provides an electronic device, as shown in fig. 3, and fig. 3 is a schematic structural diagram of the electronic device provided in the embodiment of the present invention. Comprises a processor 301, a communication interface 302, a memory 303 and a communication bus 304, wherein the processor 301, the communication interface 302 and the memory 303 are communicated with each other through the communication bus 304,
a memory 303 for storing a computer program;
the processor 301, when executing the program stored in the memory 303, implements the following steps:
acquiring play related parameters of a plurality of users aiming at the candidate video in a preset historical period; the plurality of users includes a first user and a second user; the candidate videos include a first video and a second video; the first user is a user watching a first video but not watching a second video; the second user is a user watching the first video and the second video;
determining the target preference degree of each user to the viewed candidate video according to the playing related parameters; the target preference degree is in direct proportion to the playing related parameter;
determining the target similarity of the first user and the second user according to the target preference degree of the first user to the first video and the target preference degree of the second user to the first video; the target similarity is used for representing the similarity degree between the first user and the second user;
determining the predicted preference degree of the first user for the second video according to the target similarity of the first user and a neighbor user, the first average target preference degree of the first user, the second average target preference degree of the neighbor user and the target preference degree of the neighbor user for the second video; the neighbor user is a second user of which the target similarity with the first user is greater than a first preset threshold; the first average target like degree is an average value of the target like degrees of the first user to the first video; the second average target like degree of the neighbor user is an average value of the target like degrees of the neighbor user to the first video and the second video;
determining a target video in the second video based on the predicted like-degree, and recommending the target video to the first user.
The communication bus mentioned in the above terminal may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The communication bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one thick line is shown, but this does not mean that there is only one bus or one type of bus.
The communication interface is used for communication between the terminal and other equipment.
The Memory may include a Random Access Memory (RAM) or a non-volatile Memory (non-volatile Memory), such as at least one disk Memory. Optionally, the memory may also be at least one memory device located remotely from the processor.
The Processor may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like; the Integrated Circuit may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component.
In yet another embodiment of the present invention, a computer-readable storage medium is further provided, which has instructions stored therein, and when the instructions are executed on a computer, the instructions cause the computer to execute the video recommendation method according to any one of the above embodiments.
In yet another embodiment of the present invention, there is also provided a computer program product containing instructions which, when run on a computer, cause the computer to perform the video recommendation method of any of the above embodiments.
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, cause the processes or functions described in accordance with the embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, server, or data center to another website site, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that incorporates one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.
It is noted that, herein, relational terms such as first and second, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Also, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in a process, method, article, or apparatus that comprises the element.
All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.

Claims (10)

1. A method for video recommendation, comprising:
acquiring play related parameters of a plurality of users aiming at the candidate video in a preset historical period; the plurality of users includes a first user and a second user; the candidate videos include a first video and a second video; the first user is a user watching a first video but not watching a second video; the second user is a user watching the first video and the second video;
determining the target preference degree of each user to the viewed candidate video according to the playing related parameters; the target preference degree is in direct proportion to the playing related parameter;
determining the target similarity of the first user and the second user according to the target preference degree of the first user to the first video and the target preference degree of the second user to the first video; the target similarity is used for representing the similarity degree between the first user and the second user;
determining the predicted preference degree of the first user for the second video according to the target similarity of the first user and a neighbor user, the first average target preference degree of the first user, the second average target preference degree of the neighbor user and the target preference degree of the neighbor user for the second video; the neighbor user is a second user of which the target similarity with the first user is greater than a first preset threshold; the first average target like degree is an average value of the target like degrees of the first user to the first video; the second average target like degree of the neighbor user is an average value of the target like degrees of the neighbor user to the first video and the second video;
determining a target video in the second video based on the predicted like-degree, and recommending the target video to the first user.
2. The method according to claim 1, wherein the obtaining of the play-related parameters of the candidate videos by the plurality of users in the preset history period comprises:
acquiring play-related behavior logs of the users, and writing the play-related behavior logs into a preset database;
screening out playing related parameters of the plurality of users aiming at the candidate videos from the playing related behavior log according to a preset historical sub-period; the preset history period consists of a plurality of continuous preset history sub-periods;
and carrying out aggregation statistics on the playing related parameters in the preset history sub-periods according to the preset history period to obtain the playing related parameters in the preset history period.
3. The method of claim 2, wherein before the filtering out the play-related parameters of the plurality of users for the candidate video from the play-related behavior log according to the preset history sub-period, the method further comprises:
determining a target thermal video and a video identifier of the target thermal video; the target hot video is a video with the average playing time length in at least one preset history period being greater than a second preset threshold value;
and filtering the play-related behavior log related to the target hot video from the play-related behavior log based on the video identification of the target hot video.
4. The method of claim 1, wherein prior to said determining the predicted like-degree of the first user for the second video, the method further comprises:
for each candidate video, determining the average value of the target preference degrees of all users who watch the candidate video for the candidate video to obtain the average preference degree;
and in the case that the average degree of preference is lower than a third preset threshold value, determining the candidate video as a target cold video and deleting the target cold video from the candidate video.
5. The method of claim 1, wherein determining the predicted like-degree of the first user for the second video comprises:
screening a second user with the target similarity larger than a first preset threshold value with the first user as the neighbor user based on the target similarity, and adding the neighbor user to a neighbor user set N;
for each second video, inputting the target similarity of the first user and a neighbor user, the first average target like degree, the second average target like degree of the neighbor user and the target like degree of the neighbor user to the second video into a target prediction function, and determining the predicted like degree of the first user to the second video;
wherein the target prediction function is:
Figure FDA0003199607230000031
wherein p isu,iIn order to predict the preference degree, i is a second video to be predicted; u is a first user, and N is a neighbor user in the neighbor user set N; s (u, n) is the target similarity of the first user u and the neighbor user n;
Figure FDA0003199607230000032
the first average target preference level;
Figure FDA0003199607230000033
the second average target preference level; r isn,iA target like-degree of the second video i to be predicted for the neighboring user n.
6. The method of claim 1, wherein determining a target video in the second video based on the predicted like-degree comprises:
according to the predicted preference degree, sorting the second video in a descending order according to the predicted preference degree;
and selecting a preset number of second videos sequenced in the front as the target videos according to the sequencing sequence.
7. A video recommendation apparatus, comprising:
the acquisition module is used for acquiring play related parameters of a plurality of users aiming at the candidate video in a preset history period; the plurality of users includes a first user and a second user; the candidate videos include a first video and a second video; the first user is a user watching a first video but not watching a second video; the second user is a user watching the first video and the second video;
a first determining module, configured to determine, according to the play-related parameter, a target preference degree of each user for the viewed candidate video; the target preference degree is in direct proportion to the playing related parameter;
a second determining module, configured to determine a target similarity between the first user and the second user according to a target like degree of the first user to the first video and a target like degree of the second user to the first video; the target similarity is used for representing the similarity degree between the first user and the second user;
a third determining module, configured to determine a predicted preference degree of the first user for the second video according to the target similarity between the first user and a neighboring user, the first average target preference degree of the first user, the second average target preference degree of the neighboring user, and the target preference degree of the neighboring user for the second video; the neighbor user is a second user of which the target similarity with the first user is greater than a first preset threshold; the first average target like degree is an average value of the target like degrees of the first user to the first video; the second average target like degree of the neighbor user is an average value of the target like degrees of the neighbor user to the first video and the second video;
and the recommending module is used for determining a target video in the second video based on the predicted preference degree and recommending the target video to the first user.
8. The apparatus of claim 7, wherein the obtaining module is specifically configured to:
acquiring play-related behavior logs of the users, and writing the play-related behavior logs into a preset database;
screening out playing related parameters of the plurality of users aiming at the candidate videos from the playing related behavior log according to a preset historical sub-period; the preset history period consists of a plurality of continuous preset history sub-periods;
and carrying out aggregation statistics on the playing related parameters in the preset history sub-periods according to the preset history period to obtain the playing related parameters in the preset history period.
9. An electronic device is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor and the communication interface are used for realizing mutual communication by the memory through the communication bus;
a memory for storing a computer program;
a processor for implementing the video recommendation method steps of any of claims 1-6 when executing a program stored in a memory.
10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out a video recommendation method according to any one of claims 1-6.
CN202110900488.3A 2021-08-06 2021-08-06 Video recommendation method and device, electronic equipment and readable storage medium Pending CN113626698A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110900488.3A CN113626698A (en) 2021-08-06 2021-08-06 Video recommendation method and device, electronic equipment and readable storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110900488.3A CN113626698A (en) 2021-08-06 2021-08-06 Video recommendation method and device, electronic equipment and readable storage medium

Publications (1)

Publication Number Publication Date
CN113626698A true CN113626698A (en) 2021-11-09

Family

ID=78383061

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110900488.3A Pending CN113626698A (en) 2021-08-06 2021-08-06 Video recommendation method and device, electronic equipment and readable storage medium

Country Status (1)

Country Link
CN (1) CN113626698A (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103209342A (en) * 2013-04-01 2013-07-17 电子科技大学 Collaborative filtering recommendation method introducing video popularity and user interest change
WO2017161744A1 (en) * 2016-03-25 2017-09-28 乐视控股（北京）有限公司 Video recommendation method and apparatus
CN109284417A (en) * 2018-08-27 2019-01-29 广州飞磨科技有限公司 Video pushing method, device, computer equipment and storage medium
US20190253760A1 (en) * 2018-02-12 2019-08-15 Beijing Baidu Netcom Science And Technology Co., Ltd. Method and apparatus for recommending video

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103209342A (en) * 2013-04-01 2013-07-17 电子科技大学 Collaborative filtering recommendation method introducing video popularity and user interest change
WO2017161744A1 (en) * 2016-03-25 2017-09-28 乐视控股（北京）有限公司 Video recommendation method and apparatus
US20190253760A1 (en) * 2018-02-12 2019-08-15 Beijing Baidu Netcom Science And Technology Co., Ltd. Method and apparatus for recommending video
CN109284417A (en) * 2018-08-27 2019-01-29 广州飞磨科技有限公司 Video pushing method, device, computer equipment and storage medium

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
俞美华;: "融合用户兴趣度与项目相关度的电影推荐算法研究", 电脑知识与技术, no. 08, pages 22 - 26 *
王丽娜;张学恒;王伟晨;: "基于协同过滤算法的智能推荐系统研究", 辽宁工业大学学报(社会科学版), no. 03 *

Similar Documents

Publication Publication Date Title
CN108520076B (en) Electronic book recommendation method, electronic device and computer storage medium
CN110659388B (en) To-be-recommended information screening method and device, electronic equipment and storage medium
CN108304512B (en) Video search engine coarse sorting method and device and electronic equipment
CN110149540B (en) Recommendation processing method and device for multimedia resources, terminal and readable medium
US8818928B2 (en) Evaluating an item based on user reputation information
CN108885624B (en) Information recommendation system and method
CN110929052A (en) Multimedia resource recommendation method and device, electronic equipment and storage medium
CN109903086B (en) Similar crowd expansion method and device and electronic equipment
CN112328906A (en) Content item recommendation method, device, equipment and storage medium
CN111159563A (en) Method, device and equipment for determining user interest point information and storage medium
US10733244B2 (en) Data retrieval system
CN111327609B (en) Data auditing method and device
CN115455280A (en) Recommendation list determining method and server
CN114428901A (en) Personalized data loading method and device, electronic equipment and storage medium
CN113873330A (en) Video recommendation method and device, computer equipment and storage medium
CN111858245A (en) Abnormal data analysis method and device, electronic equipment and storage medium
CN111314790A (en) Video playing record sequencing method and device and electronic equipment
JP5813052B2 (en) Information processing apparatus, method, and program
CN111177561B (en) Test method and device
CN112995765A (en) Network resource display method and device
CN113220974A (en) Click rate prediction model training and search recall method, device, equipment and medium
CN111666481B (en) Data mining method, device, computer readable medium and electronic equipment
CN111460181A (en) Multimedia resource recommendation method, electronic device and storage medium
CN113076487B (en) User interest characterization and content recommendation method, device and equipment
CN111353052B (en) Multimedia object recommendation method and device, electronic equipment and storage medium

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