CN116431841A

CN116431841A - Method and device for determining media information, electronic equipment and storage medium

Info

Publication number: CN116431841A
Application number: CN202210004812.8A
Authority: CN
Inventors: 李聪; 王震; 翼升
Original assignee: Beijing Dajia Internet Information Technology Co Ltd
Current assignee: Beijing Dajia Internet Information Technology Co Ltd
Priority date: 2022-01-04
Filing date: 2022-01-04
Publication date: 2023-07-14

Abstract

The disclosure relates to a method and a device for determining media information, an electronic device and a storage medium, wherein the method comprises the following steps: the method comprises the steps of obtaining a similarity set associated with a media information set, determining a first play record of a first group of media information played by a target account, and determining a second group of media information according to the similarity set and the first play record, wherein the second group of media information is media information to be recommended to the target account in the media information set, further, the damage rate of the second group of media information in a rough arrangement stage is reduced, personalized interest preference of a user is consulted, the diversity of a final recommendation result is increased, the overall recommendation experience of the media information is improved, and the technical problem that single interest points in the recommendation result exist in the related technology and personalized requirements of the user cannot be met is solved.

Description

Method and device for determining media information, electronic equipment and storage medium

Technical Field

The present disclosure relates to the field of computers, and in particular, to a method and apparatus for determining media information, an electronic device, and a storage medium.

Background

In the prior art, under the general condition, the main aim of media information waterfall stream recommendation is to improve the consumption experience of users, namely, improve the indexes of watching, praying, paying attention to and the like of the users, the conventional scattering strategy in the industry is mostly concentrated in the fine ranking and rearrangement stage in the whole recommendation process, the corresponding candidate set is more limited, the final recommendation result is more likely to be converged on the strongest single interest point of the users, so that the problems of single recommendation result, easy aesthetic fatigue and the like of the users are caused.

BRIEF SUMMARY OF THE PRESENT DISCLOSURE

The disclosure provides a method and a device for determining media information, electronic equipment and a storage medium, which are used for at least solving the technical problem that single interest points in recommendation results exist in related technologies and cannot meet personalized requirements of users. The technical scheme of the present disclosure is as follows:

according to a first aspect of an embodiment of the present disclosure, there is provided a method for determining media information, including:

acquiring a similarity set associated with a media information set, wherein the media information set is a set formed by media information of which a group of accounts play in a historical time period, the media information set at least comprises two different types of media information, the similarity set is formed by the similarity between the media information of each type, and the similarity is used for indicating the association degree between the media information of any two types;

determining a first play record of a first group of media information played by a target account, wherein the group of accounts comprises the target account, and the first play record is a play record in the historical time period;

and determining a second group of media information according to the similarity set and the first play record, wherein the second group of media information is media information to be recommended to the target account in the media information set.

In an exemplary embodiment, the determining the first play record of the target account to play the first set of media information includes:

acquiring the playing times of the media information in the first group of media information, wherein the playing times comprise the times of playing each type of media information in the historical time period by the target account;

and determining the first play record according to the play times.

In an exemplary embodiment, the obtaining a similarity set associated with a media information set includes:

acquiring a second play record of the media information set in the historical time period, wherein the second play record comprises play records of all accounts in the group of accounts;

and determining the similarity according to the second play record to obtain the similarity set.

In an exemplary embodiment, the determining the similarity according to the second play record, to obtain the similarity set, includes:

determining a first account set which is watched with the first type of media information and a second account set which is watched with the second type of media information according to the second play record, wherein the group of accounts comprises the first account set and the second account set;

According to the second play record, determining the account number from which the first-class media information is watched as a third account number set;

and determining the similarity between the first-type media information and the second-type media information according to the first account set, the second account set and the third account set, wherein the similarity set at least comprises the similarity between the first-type media information and the second-type media information.

In an exemplary embodiment, the determining the similarity between the first type of media information and the second type of media information according to the first set of accounts, the second set of accounts, and the third set of accounts includes:

determining the union of the first account set and the second account set as a fourth account set;

and determining the ratio of the number of the accounts of the third account set to the number of the accounts of the fourth account set as the similarity between the first-class media information and the second-class media information.

In an exemplary embodiment, the determining the second set of media information according to the similarity set and the first play record includes:

Determining a target preference vector of the target account for each media information in the first group of media information according to the similarity set and the first play record, wherein the dimension of the target preference vector corresponds to the category of the media information in the first group of media information, and the value of the target preference vector on each dimension represents the preference degree of the target account for the media information of the category corresponding to each dimension;

and determining a second group of media information to be recommended according to the target preference vector.

In an exemplary embodiment, the determining, according to the similarity set and the first play record, the target preference vector of the target account for each media information in the first set of media information includes:

generating a target similarity matrix according to the similarity set, wherein the target similarity matrix is used for representing the similarity between various categories of media information in the media information set;

determining target parameters of the target account for each media information in the first group of media information according to the first play record and the target similarity matrix, wherein the target parameters are used for representing preference degrees of the target account for each media information in the first group of media information in the historical time period;

Acquiring comprehensive preference vectors of the group of accounts for all media information in the first group of media information, wherein the comprehensive preference vectors are used for representing preference degrees of the group of accounts for all media information in the first group of media information in the historical time period;

and determining the target preference vector according to the comprehensive preference vector and the target parameter.

In an exemplary embodiment, the determining, according to the target preference vector, the second set of media information to be recommended from the media information set includes:

the method comprises the steps of obtaining ordering parameters of all media information in a media information set, wherein the ordering parameters are used for ordering all media information in the media information set according to the size of the ordering parameters in a coarse ordering stage;

and determining the second group of media information according to the sorting parameters and the target preference vector.

In an exemplary embodiment, the determining the second set of media information according to the ranking parameter and the target preference vector includes:

acquiring the estimated number N of media information needing to enter a fine-ranking stage, wherein N is a positive integer greater than 1;

Under the condition that the number of media information in the media information set is larger than N, determining the media information with the top K bits of ranking parameters in the media information set as a first media information subset, wherein K is a positive integer smaller than N;

determining a second media information subset from the media information set according to the sequence from high to low of the sorting parameters, wherein the proportion of each type of media information in the second media information subset is determined by the target preference vector, and the number of the media information in the second media information subset is N-K;

the first subset of media information and the second subset of media information are determined to be the second set of media information.

and determining the second group of media information from the media information set according to the sequence from high to low of the sorting parameters, wherein the proportion of each type of media information in the second group of media information is determined by the target preference vector, and the number of the media information in the second group of media information is N.

According to a second aspect of the embodiments of the present disclosure, there is provided a determining apparatus of media information, including:

the acquisition module is used for acquiring a similarity set associated with the media information set, wherein the media information set is a set formed by media information of which a group of accounts play in a historical time period, the media information set at least comprises two different categories of media information, the similarity set is formed by similarity among the media information of each category, and the similarity is used for indicating the association degree between the media information of any two categories;

a first determining module, configured to determine a first play record of a target account for playing a first set of media information, where the set of accounts includes the target account, and the first play record is a play record in the historical time period;

and the second determining module is used for determining a second group of media information according to the similarity set and the first play record, wherein the second group of media information is media information to be recommended to the target account in the media information set.

In an exemplary embodiment, the apparatus is configured to determine a first play record of the target account playing the first set of media information by:

Acquiring the playing times of the media information of each category in the first group of media information, wherein the playing times comprise the times of the target account playing the media information of each category in the historical time period;

and determining the first play record according to the play times.

In an exemplary embodiment, the apparatus is configured to obtain a set of similarities associated with a set of media information by:

In an exemplary embodiment, the apparatus is configured to determine the similarity from the second play record, to obtain the set of similarities, by:

In an exemplary embodiment, the apparatus is configured to determine a similarity between the first type of media information and the second type of media information according to the first set of accounts, the second set of accounts, and the third set of accounts by:

In an exemplary embodiment, the apparatus is configured to determine a second set of media information from the set of similarities and the first play record by:

In an exemplary embodiment, the apparatus is configured to determine, according to the similarity set and the first play record, a target preference vector of the target account for each media information in the first set of media information by:

In an exemplary embodiment, the apparatus is configured to determine a second set of media information to be recommended from the set of media information according to the target preference vector by:

In an exemplary embodiment, the apparatus is configured to determine the second set of media information according to the ranking parameter and the target preference vector by:

According to still another aspect of the embodiments of the present disclosure, there is provided an electronic device for determining media information, including: a processor; a memory for storing the processor-executable instructions; wherein the processor is configured to execute the instructions to implement the method of determining media information as described above.

According to yet another aspect of the embodiments of the present disclosure, there is also provided a computer-readable storage medium having a computer program stored therein, wherein the computer program is configured to perform the above-described method of determining media information when run.

The technical scheme provided by the embodiment of the disclosure at least brings the following beneficial effects: the method comprises the steps of obtaining a similarity set associated with a media information set, wherein the media information set is a set formed by media information of which a group of accounts play in a historical time period, the media information set at least comprises two different types of media information, the similarity set is formed by similarity among the media information of each type, the similarity is used for indicating the association degree between the media information of any two types, a first play record of a first group of media information is determined by a target account, one group of accounts comprises the target account, the first play record is a play record in the historical time period, and a second group of media information is determined according to the similarity set and the first play record, wherein the second group of media information is media information to be recommended to the target account in the media information set, so that the damage rate of the second group of media information in a coarse arrangement stage is reduced, personalized interest preference of a user is referred, the diversity of a final recommendation result is increased, the overall recommendation experience of the media information is improved, and the technical problem that the single interest point in the recommendation result in the related technology cannot meet the personalized demand is solved.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the disclosure.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the disclosure and together with the description, serve to explain the principles of the disclosure and do not constitute an undue limitation on the disclosure.

FIG. 1 is a schematic illustration of an application environment for a method of determining media information, according to an exemplary embodiment;

FIG. 2 is a flow chart illustrating a method of determining media information according to an exemplary embodiment;

FIG. 3 is a schematic diagram illustrating a method of determining media information according to an exemplary embodiment;

FIG. 4 is a matrix diagram illustrating a method of determining media information according to an exemplary embodiment;

FIG. 5 is a flowchart illustrating another method of determining media information according to an exemplary embodiment;

FIG. 6 is a block diagram of a media information determination apparatus according to an exemplary embodiment;

fig. 7 illustrates a block diagram of an electronic device, according to an example embodiment.

Detailed Description

In order to enable those skilled in the art to better understand the technical solutions of the present disclosure, the technical solutions of the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings.

It should be noted that the terms "first," "second," and the like in the description and claims of the present disclosure and in the foregoing figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments of the disclosure described herein may be capable of operation in sequences other than those illustrated or described herein. The implementations described in the following exemplary examples are not representative of all implementations consistent with the present disclosure. Rather, they are merely examples of apparatus and methods consistent with some aspects of the present disclosure as detailed in the accompanying claims.

First, partial terms or terminology appearing in describing embodiments of the present application are applicable to the following explanation:

the main flow of the recommendation system is mainly divided into four stages of recall, coarse ranking, fine ranking and rearrangement, wherein the recall stage is mainly used for recall of users meeting specific conditions, the coarse ranking is used for overall performance consideration, relatively simplified ranking models and formulas are used for roughly screening candidate sets before the user enters the fine ranking, the number of the required fine ranking models is reduced, the final benefit effect is ensured, the final ranking and rearrangement are used for carrying out final ranking on the candidate sets by using more complex models and strategies, and the consumption experience of the user is improved to the greatest extent.

The invention is illustrated below with reference to examples:

according to an aspect of the embodiment of the present invention, there is provided a method for determining media information, optionally, in this embodiment, the method for determining media information described above may be applied to a hardware environment composed of the server 101 and the terminal device 103 as shown in fig. 1. As shown in fig. 1, a server 101 is connected to a terminal 103 through a network, and may be used to provide services to a user terminal or a client installed on the user terminal, which may be a video client, an instant messaging client, a browser client, an educational client, a game client, etc. The database 105 may be provided on or independent of the server for providing data storage services for the server 101, such as a media information data storage server, which may include, but is not limited to: a wired network, a wireless network, wherein the wired network comprises: local area networks, metropolitan area networks, and wide area networks, the wireless network comprising: bluetooth, WIFI and other wireless communication networks, the terminal device 103 may be a terminal configured with an application program, and may include, but is not limited to, at least one of the following: the mobile phone (such as an Android mobile phone, an iOS mobile phone, etc.), a notebook computer, a tablet computer, a palm computer, an MID (Mobile Internet Devices, mobile internet device), a PAD, a desktop computer, a smart television, etc., where the server may be a single server, a server cluster formed by a plurality of servers, or a cloud server, and the application program 107 using the method for determining media information is displayed through a terminal device 103 or a client installed on the server 101.

As shown in fig. 1, the above method for determining media information may be implemented at the server 101 by:

s1, acquiring a similarity set associated with a media information set on a server 101, wherein the media information set is a set formed by media information played by a group of accounts in a historical time period, the media information set at least comprises two different types of media information, the similarity set is formed by similarities among the media information of all the types, and the similarities are used for indicating the association degree between the media information of any two types;

s2, determining a first play record of a first group of media information played by the target account on the server 101, wherein the first play record is a play record in a historical time period;

s3, determining a second group of media information according to the similarity set and the first play record on the server 101, wherein the second group of media information is media information to be recommended to the target account in the media information set.

Alternatively, in the present embodiment, the above-described method of determining media information may be implemented on the terminal device 103 in a similar manner, or may be implemented asynchronously by the server 101 and the terminal device 103.

The above is merely an example, and the present embodiment is not particularly limited.

Optionally, as an optional implementation manner, as shown in fig. 2, the method for determining media information includes:

s202, acquiring a similarity set associated with a media information set, wherein the media information set is a set formed by media information of which a group of accounts play in a historical time period, the media information set at least comprises media information of two different categories, the similarity set is formed by similarity among media information of each category, and the similarity is used for indicating the association degree between the media information of any two categories;

s204, determining a first play record of the target account for playing the first group of media information, wherein one group of accounts comprises the target account, and the first play record is a play record in a historical time period;

s206, determining a second group of media information according to the similarity set and the first play record, wherein the second group of media information is media information to be recommended to the target account in the media information set.

Optionally, in this embodiment, the method for determining media information may be applied to determining media information in a plurality of application scenarios including medical treatment, finance, credit investigation, banking, energy, education, security, building, game, traffic, internet of things, industry, and the like, where media information is recommended to a terminal device through the method for determining media information.

Alternatively, in this embodiment, the media information set may include, but is not limited to, a recall set obtained in a recall phase, and the media information may include, but is not limited to, media information that needs to be recommended to the user, such as an advertisement, a short video, a topic, and the like.

For example, fig. 3 is a schematic diagram of a method for determining media information according to an exemplary embodiment, as shown in fig. 3, where a recall process of media information includes a recall stage, a coarse ranking node, a fine ranking node, and a rearrangement stage, and the embodiment of the present application may include, but is not limited to, a coarse ranking stage, where in a general case, a main recommendation goal of media information is to promote a user's consumption experience, that is, promote indexes such as watching, praying, and paying attention, etc. of a user, a conventional scattering policy in industry concentrates on a fine ranking and a rearrangement stage in a plurality of points in the overall recommendation process, a corresponding candidate set is limited, a final recommendation result is likely to converge to a single point of interest where a user is strongest, thereby causing a problem of single recommendation result, easy aesthetic fatigue, etc., and on the other hand, a conventional category scattering is concentrated on a regular constraint, statistical information of the same city user with respect to a media information category preference is not considered, and a break-down effect on media information is relatively large.

Alternatively, in the present embodiment, the above-described history time period may be preset by the system, for example, configured to be 1 week or 1 month or the like.

Optionally, in this embodiment, the media information set includes at least two different types of media information, which may include, but is not limited to, at least two types of media information including medical, financial, credit, banking, energy, education, security, building, game, traffic, internet of things, industry, and the like.

For example, the media information 1 is a financial type media information, the media information 2 is an educational type media information, and the media information 3 is a game type media information.

Optionally, in this embodiment, the similarity set between the media information of each category in the media information set may include, but is not limited to, jacard similarity, and the interest preference of the user is calculated by using the Jacard similarity between the categories, so that the interest preference of the user can be accurately calculated through the play record of the historical time period under the condition that the user behavior is rare.

For example, the account a plays the media information 1 in the history time period, the account B plays the media information 2, the media information 3, and the account C plays the

media information

1, 3 in the history time period, the media information 1 is the financial type media information, the media information 2 is the government type media information, the media information 3 is the game type media information, and the similarity between the type corresponding to the media information 1 and the type corresponding to the media information 3 is 1/4, which is calculated as follows:

Wherein the Similarity is _AB The similarity between the media information of the A category and the media information of the B category is represented, A and B represent the total number of users simultaneously playing the media information of the A category and the media information of the B category, and A and B represent the total number of users playing the media information of the A category and the media information of the B category.

Optionally, in this embodiment, the above-mentioned collection of playing times of playing each type of media information in the first set of media information by the target account in the history time period may include, but is not limited to, a collection of playing times of playing each media information in the first set of media information by the target account in the history time period.

For example, the first group of media information includes media information 1, media information 2 and media information 3, the target account plays media information 1 once, plays media information 2 three times and does not play media information 3 in the historical time period, the play times set is that the number of times of playing media information corresponding to the category of media information 1 of the target account is 1, the number of times of playing media information corresponding to the category of media information 2 is 3, that is, the target account prefers to play media information corresponding to the category of media information 2, and secondly media information corresponding to the category of media information 1, and the target account does not like media information corresponding to the category of media information 3.

Optionally, in this embodiment, the first play record may include, but is not limited to, a number of clicks, a number of slides, a number of voice triggers, a number of gesture triggers, and so on the media information, and specifically may include, but is not limited to, a number of successful presentations of the media information, or a number of successful presentations completed.

Optionally, in this embodiment, determining the second set of media information according to the similarity set and the first play record may include, but is not limited to, determining the second set of media information according to the similarity set and the first play record based on a result of the coarse-ranking stage in a coarse-ranking stage of the media information set, so as to send the second set of media information to a fine-ranking stage, thereby completing a subsequent fine-ranking, rearrangement, and other processes.

According to the embodiment, a similarity set associated with a media information set is obtained, wherein the media information set is a set formed by media information of which a group of accounts play in a historical time period, the media information set at least comprises two different types of media information, the similarity set is formed by similarity between the media information of each type, the similarity is used for indicating the association degree between the media information of any two types, a first play record of a first group of media information is determined by a target account, one group of accounts comprises the target account, the first play record is a play record in the historical time period, and a second group of media information is determined according to the similarity set and the first play record, wherein the second group of media information is the media information to be recommended to the target account in the media information set, so that the breakage rate of the second group of media information in a coarse ranking stage is reduced, the personalized interest preference of a user is referred, the diversity of a final recommendation result is increased, the overall recommendation of the media information is promoted, and the technical problem that the interest point in the recommendation result is single in the related technology cannot meet the personalized requirements of the user is solved.

and determining the first play record according to the play times.

Alternatively, in this embodiment, the number of playing times may include, but is not limited to, the number of playing times of the media information played by the target account in the historical time period, where the number of playing times is counted according to the type, for example. The number of times that the target account plays the media information A is 1, the number of times that the media information B is 2, the number of times that the media information C is 3, wherein the media information A and the media information B both belong to the media information of the type X, the media information C belongs to the media information of the type Y, and the playing times are the playing times 3 of the media information of the type X and the playing times 3 of the media information of the type Y.

Alternatively, in this embodiment, the determining the first play record according to the number of play times may include, but is not limited to, determining the number of play times of each type of media information played by the target account as the first play record.

According to the embodiment, the playing record of the target account on the historical time period is obtained, so that the subsequently generated second group of media information is the media information obtained by combining the playing habit of the target account on the historical time period, the consumption effect of the user is further ensured, the diversity of the final recommendation result is increased, and the recommendation experience and sustainable consumption of the media information application are improved.

Optionally, in this embodiment, the second play record includes a set of accounts playing media information and a media information history record played by each account in the set of accounts.

For example, the second play record includes a group of accounts including an account a, an account B, and an account C for playing media information, where the account a plays media information 1 in a history time period, the account B plays

media information

2 and 3 in the history time period, and the account C plays

media information

1 and 3 in the history time period.

According to the embodiment, through the play record of a group of accounts in the historical time period, the interest preference of the user can be accurately calculated through the large disc information of the media information recommendation application, and by taking the example that a group of accounts are the same city accounts, the media information recommended for the target account can be adjusted according to the watching habit of the same city accounts, so that the technical effect of enriching the diversity of recommendation results is achieved.

Alternatively, in this embodiment, the first type of media information may include, but is not limited to, any type of media information in the media information set, and the second type of media information may include, but is not limited to, media information in a different type from the first type of media information.

Optionally, in this embodiment, the accounts in the first account set are accounts that have seen the first type of media information in the historical time period, the accounts in the second account set are accounts that have seen the second type of media information in the historical time period, and the accounts in the third account set are accounts that have seen both the first type of media information and the second type of media information in the historical time period.

Optionally, in this embodiment, the similarity between the media information of each category is calculated for the media information set, so as to obtain a similarity set.

According to the embodiment, the first account set, the second account set and the third account set are determined according to the second play record, and then the similarity between the first type media information and the second type media information is determined according to the first second account set and the third account set, so that the interest preference of the user can be accurately calculated through the large disc information (the second play record) under the condition that the user behavior corresponding to the target account is rare, the diversity of the final recommendation result is increased, the overall recommendation experience of the media information is improved, and the technical problem that the interest point in the recommendation result is single and the personalized requirement of the user cannot be met in the related technology is solved.

Specifically, it may include, but is not limited to:

s1, acquiring the first account number of a first account set, the second account number of a second account set and the third account number of a third account set;

s2, determining the sum value of the number of the first accounts and the number of the second accounts;

s3, determining the ratio of the number of the third accounts to the sum value as the similarity between the first-class media information and the second-class media information.

Optionally, in this embodiment, the sum of the number of the first account and the number of the second account may be subjected to the deduplication processing, and when a certain account views both the first type of media information and the second type of media information, statistics is performed only once.

Alternatively, in the present embodiment, the similarity between the first type of media information and the second type of media information may include, but is not limited to, a determination by the following formula:

wherein A represents first-class media information, B represents second-class media information, and Similarity _AB Representing the similarity between the first and second type of media information, A n B represents the number of accounts that have been viewed both A and B, i.e., the third number of accounts, A n B represents the number of accounts that have been viewed A or B, i.e., the sum of the first and second numbers of accounts.

According to the embodiment, the number of accounts for watching the first-class media information and the number of accounts for watching the second-class media information are counted respectively, and the number of accounts for watching the first-class media information and the second-class media information is also counted.

Optionally, in this embodiment, the dimensions of the target preference vector represent the number of categories of media information in the first set of media information, and the value in each dimension of the target preference vector represents the preference degree of the target account to the media information of the category corresponding to each dimension.

For example, the target preference vector is (1,2,1,4,0,4), it may be understood that the target account views 6 types of media information in the historical time period, and the preference degrees of the 6 types of media information are respectively 1,2,1,4,0 and 4, where the larger the value in each dimension, the more preferred the target account is for the corresponding type of media information in the dimension.

According to the embodiment, the clicking behaviors of the target account on the latest categories are utilized to calculate the personalized category preference of the user, the category preference is represented by a category preference vector, the dimension of the vector is the classification number of the current short video category, the value on each dimension of the vector represents the preference classification of the user on the category, the personalized interest preference of the user is considered, diversified recommendation results of the user are enriched, the overall recommendation efficiency of media information is improved, and the technical problem that the single interest point in the recommendation results exists in the related technology and the personalized requirements of the user cannot be met is solved.

Optionally, in this embodiment, the generating the target similarity matrix according to the similarity set may include, but is not limited to, generating the target similarity matrix according to a similarity between media information in each category in the similarity set, where each row and each column of the target similarity matrix corresponds to a category of media information, and an autonomy of media information in each category is configured to be 1.

Fig. 4 is a schematic diagram of a similarity matrix of a method for determining media information according to an exemplary embodiment, as shown in fig. 4, where the similarity matrix is a matrix of N rows and N columns, and 1 to 5 represent media information of 5 categories included in a media information set, and then similarity between the media information of 5 categories is represented by the matrix.

Optionally, in this embodiment, determining, according to the play number set and the target similarity matrix, the target parameter of the target account for each media information in the first group of media information may include, but is not limited to, multiplying the number of clicks of the target account for each media information in the first group of media information by the similarity matrix, and summing the multiplied number of clicks of each media information in the first group of media information to obtain the target parameter.

Specifically, it may include, but is not limited to, one obtained by the following formula:

Score _i ＝CatClick _i +∑(Similarity _ij *CatClick _j )

wherein Score _i Target parameters, catClick, representing media information of category i _i Representing the number of plays of the target account for the i-category media information, similarity _ij Representing the target similarity matrix.

Optionally, in this embodiment, the obtaining, from the play record, the integrated preference vector of the set of accounts for each media information in the first set of media information may include, but is not limited to, obtaining, from the play record, an integrated preference vector of a set of accounts, where the integrated preference vector may include, but is not limited to, determining, according to a number of plays of each account in the set of accounts for each media information in the media information set.

Optionally, in this embodiment, determining the target preference vector according to the integrated preference vector and the target parameter may include, but is not limited to, determining by the following formula:

CatPreference _i ＝a*NormalizedScore _i +(1-a)*AvgCatPreference _i

wherein, catPreference _i Representing the value of the i category media information dimension in the target preference vector, normazedscore _i Target parameters representing normalized i-category media information, avgcatreference _i And (3) representing the value of the i-category media information in the comprehensive preference vector, wherein a represents a preset weight.

According to the embodiment, based on the statistical information of the category preference of the media information of the group of accounts associated with the comprehensive preference vector, the damage to the ordering effect of the media information is reduced, the overall recommendation efficiency of the media information is improved, and the technical problem that the single interest point in the recommendation result in the related technology cannot meet the personalized requirements of the user is solved.

Optionally, in this embodiment, the ranking parameter may include, but is not limited to, a consumption score of the media information calculated in the coarse ranking stage, where the ranking parameter is used to rank each media information in the media information set according to a value of the ranking parameter from a big order to a small order.

By the method, the device and the system, the damage caused by the ordering effect of the media information is reduced, the overall recommending efficiency of the media information is improved, and the technical problem that the single interest point in the recommending result exists in the related technology and the personalized requirement of the user cannot be met is solved.

under the condition that the number of media information in the media information set is larger than N, determining the media information with the ranking parameter ranking the top K bits in the media information set as a first media information subset, wherein K is a positive integer smaller than N;

media information included in a union of the first subset of media information and the second subset of media information is determined as the second set of media information.

Optionally, in this embodiment, the estimated number N of media information needed to enter the fine-ranking stage may be preset by the system, or may be dynamically determined according to the number of times of playing the target account in the historical time period, or may be dynamically determined according to the conversion rate and/or click rate of each media information in the media information set.

Optionally, in this embodiment, media information with a ranking parameter located in the first K bits in the media information set may be cut off first to be used as the first media information subset, and then the proportion of each type may be obtained according to the target preference vector, and N-K media information may be determined one by one from media information with a ranking parameter located in the k+1th bit according to the proportion of each type.

Specifically, it may include, but is not limited to, determining the proportions of the above types by:

CatFrCount＝CatPreference _i *AllFrCount

wherein, catFrCount represents the proportion of the i-type media information in the second media information subset, catPreferencei represents the value of the i-type media information dimension in the target preference vector, and AllFrCount represents the total proportion of the various types of media information.

According to the embodiment, the method and the device can be applied to the rough ranking process of media information recommendation of a group of accounts, and finally, the consumption effect is increased while the variety of the transmitted feed is increased, so that the win-win effect of the consumption attribute and the variety of the feed is achieved, the consumption experience and the DAU permeability of the whole user are improved obviously, the whole recommendation efficiency of the media information is optimized, and the technical problem that the single interest point in the recommendation result in the related technology cannot meet the personalized requirements of the user is solved.

Alternatively, in this embodiment, the proportion of each type may be obtained according to the target preference vector, and N pieces of media information may be determined one by one according to the proportion of each type from the media information whose ranking parameter is located at the 1 st bit.

The present application is further explained below in conjunction with specific examples:

a personalized multi-interest point ordering mechanism based on Jacard similarity is designed in a coarse ordering stage, interest preference of a user is calculated by utilizing Jacard similarity among categories, the interest preference of the user can be accurately calculated through large-disc information under the condition that the user behaviors are rare, and an interest scattering mechanism is advanced from a final rearrangement stage to the coarse ordering stage, so that category diversity entering fine ordering is ensured, and the scattering in the rearrangement stage can play an effect to a greater extent.

The method comprises the following specific steps:

1. according to the large disc exposure of the user in the history category (play record on the history time period) of the same city page, the Jacard similarity of the whole user in each category of the consumption media information of the same city page is calculated, and the Jacard similarity is taken as a measure of the similarity between the media information categories, wherein the concrete calculation mode is that for A, B two categories, the corresponding similarity is calculated as follows:

wherein the degree of autonomy of a category is defined as 1.0.

2. Calculating user personalized category preference by utilizing click behaviors of the user on the latest categories of the pages in the same city;

a. the class preference is represented by a class preference vector, the dimension of the vector is the number of classifications of the current co-city media information classes, and the value in each dimension of the vector represents the preference score of the user for the class.

b. Calculating the recent click number (the watching number of each category) of the user on each category, multiplying the click number of each category by the Jaccard similarity matrix among the categories, and summing the click numbers of the categories to obtain the click score of the user on each category:

Score _i ＝CatClick _i +∑(Similarity _ij *CatClick _j )

c. normalizing the obtained click score according to the vector, and carrying out weighted average on the click score and the class average preference of the large disk in the same city to obtain final preference vectors of each class;

CatPreference _i ＝a*NormalizedScore _i +(1-a)*AvgCatPreference _i

3. In order to ensure that the transparent data set meets the consumption requirement of the user and also meets the candidate set of category diversity, the design strategy is as follows:

a. obtaining an initial sorting score for the coarse sorting result through a coarse sorting stage sorting model and a scoring formula;

b. setting the final quantity of entering the fine row as n, and for ensuring the consumption effect preferentially, the first k photo enters the fine row according to the sequencing result, and the last n-k quantity of the fine row expected to enter on each category is calculated according to the user category preference vector in proportion.

CatFrCount＝CatPreference _i *AllFrCount

Fig. 5 is a schematic diagram illustrating a method for determining media information according to an exemplary embodiment, and as shown in fig. 5, the method generally includes the following steps:

s1, calculating the similarity among categories (categories);

s2, calculating user (target account) category preference (target preference vector);

s3, calculating the number of fine arrangement of each category (the proportion of each category);

s4, ensuring that the consumption TOPK directly enters the fine row;

s5, judging whether the number of the media information which enters the fine-ranking stage is smaller than N from the K+1st media information;

S6, when the number of the media information of the same category of the media information to be determined is smaller than N, judging whether the number of the media information entering the fine arranging stage is smaller than the number of the media information needing to enter the fine arranging stage in each category;

and S7, adding the media information into a media information set to be in the fine-ranking stage under the condition that the judgment result of the step S7 is yes, skipping the media information under the condition that the judgment result of the step S7 is no, and repeating the operation from the K+2.

If the historical behaviors of the users are sparse, the variety of the categories can be expanded according to the experience category preference distribution of the large disk, if the historical behaviors of the users are dense and single, the preference of the users to other categories can be discovered through the similarity conduction among the categories, and if the historical category behaviors of the same user are various, the recommendation meeting the consumption and meeting the diversity can be obtained through the personalized diversity coarse ranking.

It will be appreciated that in the specific embodiments of the present application, related data such as user information, user play history, etc. are referred to, and when the embodiments of the present application are applied to specific products or technologies, user permission or consent is required to be obtained, and the collection, use and processing of related data is required to comply with related laws and regulations and standards of related countries and regions.

In an exemplary embodiment, fig. 6 is a block diagram of a determination device of media information according to an exemplary embodiment. Referring to fig. 6, the apparatus includes an acquisition module 602, a first determination module 604, and a second determination module 606.

The obtaining module 602 is configured to obtain a similarity set associated with a media information set, where the media information set is a set made up of media information that is played by a group of accounts in a historical time period, and the media information set includes at least two different types of media information, and the similarity set is made up of similarities between media information of respective types, where the similarities are used to indicate a degree of association between any two types of media information;

a first determining module 604 configured to determine that a target account plays a first play record of a first set of media information, wherein the set of accounts includes the target account, the first play record being a play record within the historical time period;

a second determining module 606, configured to determine a second set of media information according to the similarity set and the first play record, where the second set of media information is media information to be recommended to the target account in the media information set.

and determining the first play record according to the play times.

The specific manner in which the various modules perform the operations in the apparatus of the above embodiments have been described in detail in connection with the embodiments of the method, and will not be described in detail herein.

According to still another aspect of the embodiment of the present invention, there is also provided an electronic device for implementing the above-mentioned method for determining media information, where the electronic device may be a terminal device or a server as shown in fig. 1. The present embodiment is described taking the electronic device as a server as an example. As shown in fig. 7, the electronic device comprises a memory 702 and a processor 704, the memory 702 storing a computer program, the processor 704 being arranged to perform the steps of any of the method embodiments described above by means of the computer program.

Alternatively, in this embodiment, the electronic device may be located in at least one network device of a plurality of network devices of the computer network.

Alternatively, in the present embodiment, the above-described processor may be configured to execute the following steps by a computer program:

s1, acquiring a similarity set associated with a media information set, wherein the media information set is a set formed by media information of which a group of accounts play in a historical time period, the media information set at least comprises media information of two different categories, the similarity set is formed by similarity among media information of each category, and the similarity is used for indicating the association degree between the media information of any two categories;

S2, determining a first play record of the target account for playing the first group of media information, wherein one group of accounts comprises the target account, and the first play record is a play record in a historical time period;

and S3, determining a second group of media information according to the similarity set and the first play record, wherein the second group of media information is media information to be recommended to the target account in the media information set.

Optionally, in this embodiment, the above processor may be further configured to execute the following steps by a computer program:

Alternatively, as will be appreciated by those skilled in the art, the structure shown in fig. 7 is merely illustrative, and the electronic device may be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a palmtop computer, a mobile internet device (Mobile Internet Devices, MID), a PAD, or other terminal devices. Fig. 7 is not limited to the structure of the electronic device and the electronic apparatus described above. For example, the electronics may also include more or fewer components (e.g., network interfaces, etc.) than shown in fig. 7, or have a different configuration than shown in fig. 7.

The memory 702 may be used to store software programs and modules, such as program instructions/modules corresponding to the method and apparatus for determining media information in the embodiment of the present invention, and the processor 704 executes the software programs and modules stored in the memory 702, thereby performing various functional applications and data processing, that is, implementing the method for determining media information described above. The memory 702 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid state memory. In some examples, the memory 702 may further include memory remotely located relative to the processor 704, which may be connected to the terminal via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof. The memory 702 may be used for storing information such as historical user data, among other things. As an example, as shown in fig. 7, the memory 702 may include, but is not limited to, a first acquisition module 602, a first determination module 604, a second acquisition module 606, and a second determination module 608 in the determination device including the media information. In addition, other module units in the above media information determining apparatus may be included, but are not limited to, and are not described in detail in this example.

Optionally, the transmission device 706 is used to receive or transmit data via a network. Specific examples of the network described above may include wired networks and wireless networks. In one example, the transmission device 706 includes a network adapter (Network Interface Controller, NIC) that may be connected to other network devices and routers via a network cable to communicate with the internet or a local area network. In one example, the transmission device 706 is a Radio Frequency (RF) module that is configured to communicate wirelessly with the internet.

In addition, the electronic device further includes: a display 708 for displaying user data of the target service; and a connection bus 710 for connecting the respective module parts in the above-described electronic device.

In other embodiments, the terminal device or the server may be a node in a distributed system, where the distributed system may be a blockchain system, and the blockchain system may be a distributed system formed by connecting the plurality of nodes through a network communication. Among them, the nodes may form a Peer-To-Peer (P2P) network, and any type of computing device, such as a server, a terminal, etc., may become a node in the blockchain system by joining the Peer-To-Peer network.

In an exemplary embodiment, a computer program product is also provided, comprising a computer program/instruction which, when executed by a processor, implements the above-described information transmission method.

Other embodiments of the disclosure will be apparent to those skilled in the art from consideration of the specification and practice of the disclosure disclosed herein. This application is intended to cover any adaptations, uses, or adaptations of the disclosure following, in general, the principles of the disclosure and including such departures from the present disclosure as come within known or customary practice within the art to which the disclosure pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the disclosure being indicated by the following claims.

It is to be understood that the present disclosure is not limited to the precise arrangements and instrumentalities shown in the drawings, and that various modifications and changes may be effected without departing from the scope thereof. The scope of the present disclosure is limited only by the appended claims.

Claims

1. A method for determining media information, comprising:

2. The method of claim 1, wherein determining a first play record for the target account to play the first set of media information comprises:

and determining the first play record according to the play times.

3. The method of claim 1, wherein the obtaining a set of similarities associated with a set of media information comprises:

4. The method of claim 3, wherein said determining said similarity from said second play record to obtain said set of similarities comprises:

5. The method of claim 3, wherein the determining a second set of media information from the set of similarities and the first play record comprises:

6. The method of claim 5, wherein the determining a second set of media information to be recommended from the set of media information based on the target preference vector comprises:

7. A device for determining media information, comprising:

8. An electronic device, comprising:

a processor;

a memory for storing the processor-executable instructions;

wherein the processor is configured to execute the instructions to implement the method of determining media information as claimed in any one of claims 1 to 6.

9. A computer readable storage medium, characterized in that instructions in the computer readable storage medium, when executed by a processor of an electronic device, enable the electronic device to perform the method of determining media information according to any one of claims 1 to 6.

10. A computer program product comprising a computer program, characterized in that the computer program, when executed by a processor, implements the method of determining media information according to any one of claims 1 to 6.