CN113098934A

CN113098934A - Content pushing method based on big data and private domain flow and social network platform

Info

Publication number: CN113098934A
Application number: CN202110310698.7A
Authority: CN
Inventors: 吴小红
Original assignee: Individual
Current assignee: Individual
Priority date: 2021-03-23
Filing date: 2021-03-23
Publication date: 2021-07-09

Abstract

The embodiment of the invention provides a content pushing method based on big data and private domain flow and a social network platform, wherein the interactive user interest characteristics of service subscription content provided by each interactive platform user in the private domain flow pool of a target platform user on the social network platform and the target user interest characteristics of service subscription content of the target platform user are obtained through analysis of a user subscription information knowledge base corresponding to the private domain flow pool of the target platform user to be subjected to content pushing, and then the content pushing is carried out on the service subscription content based on the private domain flow pool of the target platform user according to the interactive user interest characteristics and the target user interest characteristics. Therefore, interest characteristic analysis of service subscription content is carried out on the target platform user and the interaction platform user having an effective interaction relation with the target platform user from different information dimensions, and then the subscription content is pushed based on the result of the interest characteristic analysis, so that the content pushing accuracy and the content pushing effect can be effectively improved.

Description

Content pushing method based on big data and private domain flow and social network platform

Technical Field

The invention relates to the technical field of big data and content pushing, in particular to a content pushing method and a social network platform based on big data and private domain flow.

Background

With the development of big data technology and the rise of various social network platforms, the social network-based private traffic is favored by various subjects, and it is expected that business is realized or business operation is performed through the private traffic, for example, advertisement delivery or content push can be performed according to the private traffic. However, in practical applications, content push based on private traffic is simply performed, and the content push effect is greatly reduced due to the problems that the push target is not targeted and there are some "invalid" private traffic in the private traffic pool, and the like, and meanwhile, there may be some potential problems of loss of valid traffic.

Disclosure of Invention

Based on the defects of the prior art, in a first aspect, an embodiment of the present invention provides a content push method based on big data and private domain traffic, including:

the method comprises the steps of obtaining a private domain flow pool of a target platform user to be subjected to content push in a social network platform and obtaining a corresponding user subscription information knowledge base according to the private domain flow pool, wherein the user subscription information knowledge base corresponding to the private domain flow pool comprises user identification information of an interaction platform user having an effective interaction relationship with the target platform user, content identification information of service subscription content provided by the social network platform, interaction relationship description information between the user identification information and content subscription description information between the user identification information and the content identification information, the interaction relationship description information comprises the interaction relationship information between the interaction platform users, and the content subscription description information comprises historical content subscription information of the interaction platform user on the service subscription content provided by the social network platform;

acquiring interactive user interest characteristics of service subscription contents provided by the social network platform for each interactive platform user in a private flow pool of the target platform user according to interactive relationship description information between the user identification information of the interactive platform user and content subscription description information between the user identification information and the content identification information;

obtaining historical content subscription information of the target platform user, and obtaining target user interest characteristics of the target platform user on service subscription content provided by the social network platform according to the historical content subscription information;

and carrying out content push on the service subscription content provided by the social network platform based on the private domain flow pool of the target platform user according to the interactive user interest characteristics and the target user interest characteristics.

Based on the first aspect, the content push of the service subscription content provided by the social network platform based on the private domain traffic pool of the target platform user according to the interactive user interest feature and the target user interest feature includes:

according to the interactive relationship description information among the user identification information, dividing a plurality of interactive platform users of which the interactive relationship meets a preset condition into an interest group to be pushed, and taking the content label of the historical subscription content of each platform user in the interest group as the interactive user interest characteristic corresponding to the interest group;

according to the interest characteristics of each interactive user, corresponding subscription content to be pushed is matched from the social network platform, and the subscription content to be pushed is pushed to the interactive platform users in the interest group corresponding to the interest characteristics of each interactive user through a first content pushing rule predetermined based on the target platform user;

and matching corresponding subscription content to be pushed from the social network platform according to the target user interest characteristics, and pushing the subscription content to be pushed to at least one interactive platform user in the private domain flow pool through a second content pushing rule predetermined based on the target platform user.

Based on the first aspect, the content pushing the subscription content to be pushed to the interaction platform users in the interest group corresponding to the interest features of the interaction users through a first content pushing rule predetermined based on the target platform user includes:

sending corresponding subscription content to be pushed to the social network clients of the interaction platform users in the interest group corresponding to the interest characteristics of the interaction users through the social network clients of the target platform users respectively;

the content pushing the subscription content to be pushed to at least one interactive platform user in the private domain flow pool through a second content pushing rule predetermined based on the target platform user comprises:

releasing the subscription content to be pushed through a preset interactive community of the target platform user, and setting the released subscription content to be pushed on the interactive community to be visible only to the interactive platform user in the private domain flow pool; or

And acquiring interaction platform users to be pushed, the interaction relationship of which with the target platform users reaches a preset condition, according to the interaction relationship description information among the user identification information, and pushing the subscription content to be pushed to the interaction platform users to be pushed through the target platform users.

Based on the first aspect, the method further comprises:

acquiring interaction relations between different interaction platform users in the private domain flow pool and the target platform user and interaction relations between the interaction platform users to obtain a platform user interaction relation matrix;

obtaining historical subscription information of each target platform user and each interaction platform user on subscription content provided by the social network platform from the social network platform to obtain a historical subscription information matrix;

and constructing the user subscription information knowledge base according to the platform user interaction relation matrix and the historical subscription information matrix, wherein the user subscription information knowledge base is a knowledge graph.

Based on the first aspect, the obtaining, according to the interaction relationship description information between the user identification information of the interaction platform users and the content subscription description information between the user identification information and the content identification information, the interaction user interest characteristics of each interaction platform user in the private domain traffic pool of the target platform user on the service subscription content provided by the social network platform includes:

converting interactive relationship description information among user identification information of each interactive platform user into an interactive relationship characteristic matrix through a knowledge base deep learning network obtained through pre-training, and converting content subscription description information between the user identification information and content identification information into a content subscription characteristic matrix;

obtaining the interactive user interest characteristics of each interactive platform user in the private domain flow pool of the target platform user on the service subscription content provided by the social network platform according to the interactive relationship characteristic matrix and the content subscription characteristic matrix;

the obtaining of the target user interest characteristics of the target platform user for the service subscription content provided by the social network platform according to the historical content subscription information includes:

according to the subscription time of the target platform user for each piece of historical content subscription information, performing information processing on each piece of historical content subscription information to obtain a time sequence-based historical subscription characteristic matrix of the target platform user;

and acquiring target user interest characteristics of the target platform user on service subscription content provided by the social network platform according to the historical subscription characteristic matrix and the content subscription characteristic matrix.

Based on the first aspect, the obtaining, according to the historical subscription feature matrix and the content subscription feature matrix, a target user interest feature of the target platform user for service subscription content provided by the social network platform includes:

and performing fusion analysis processing on the historical subscription characteristic matrix and the content subscription characteristic matrix, and taking a subscription content label corresponding to the intersection of the historical subscription characteristic matrix and the content subscription characteristic matrix as the target user interest characteristic.

Based on the first aspect, the method further comprises:

acquiring a preset traversal node combination sequence for information traversal in a user subscription information knowledge base corresponding to the private domain flow pool, wherein the traversal node combination sequence comprises a plurality of different traversal node combinations;

sequentially extracting information in a user subscription information knowledge base corresponding to the private domain flow pool according to each traversal node combination in the traversal node combination sequence to obtain a plurality of user identification information sets corresponding to the traversal node combinations respectively and a plurality of content subscription description information sets corresponding to the traversal node combinations respectively;

the method for converting the interactive relationship description information between the user identification information of each interactive platform user into an interactive relationship feature matrix and converting the content subscription description information between the user identification information and the content identification information into a content subscription feature matrix comprises the following steps:

and respectively inputting the plurality of user identification information sets and the plurality of content subscription description information sets into the knowledge base deep learning network for information processing to obtain the interaction relation characteristic matrix and the content subscription characteristic matrix.

Based on the first aspect, the method further comprises:

acquiring a knowledge base sample, wherein the knowledge base sample comprises sample user identification information of sample interaction platform users, sample content identification information of service subscription content provided by a sample social network platform, interaction relation description information between the sample user identification information and content subscription description information between the sample user identification information and the sample content identification information, the interaction relation description information comprises interaction relation information between the sample interaction platform users, and the content subscription description information comprises historical content subscription information of the sample interaction platform users on the service subscription content provided by the sample social network platform;

converting the identification information of each sample user into a sample interaction relation characteristic matrix through the knowledge base deep learning network, and converting content subscription description information between the identification information of the sample user and the identification information of the sample content into a sample content subscription characteristic matrix;

and performing iterative training on the knowledge base deep learning network according to each sample interaction relationship characteristic matrix and each sample content subscription characteristic matrix until the network model loss function value of the knowledge base deep learning network obtained in the iterative training process reaches a preset training termination condition.

performing feature fusion on the interactive user interest features of each interactive platform user in the private domain traffic pool of the target platform user on each service subscription content provided by the social network platform and the target user interest features of the target platform user for each service subscription content provided by the social network platform to obtain the target interest features of the target platform user on the service subscription content provided by each social network platform, wherein the target interest features comprise interest degree parameters of the target platform user for each service subscription content;

according to the target interest characteristics of the target platform user on the service subscription content provided by each social network platform, performing descending arrangement according to the interest degree parameters;

according to the target interest characteristics after the descending order arrangement, selecting a preset number of target interest characteristics to match subscription content to be pushed in the social network platform;

according to the matched subscription content to be pushed, pushing the content based on the private domain flow pool of the target platform user;

the method for obtaining the target interest characteristics of the service subscription content provided by the target platform user to each social network platform by performing feature fusion on the interaction user interest characteristics of each interaction platform user in the private domain traffic pool of the target platform user to each service subscription content provided by the social network platform and the target user interest characteristics of the target platform user to each service subscription content provided by the social network platform includes:

for each service subscription content, acquiring target user interest characteristics of the target platform user for the service subscription content and interactive user interest characteristics of each interactive platform user for the service subscription content;

according to the interest degree parameters included in the interest characteristics of the interaction users of the service subscription contents by the interaction platform users, correcting the interest degree parameters included in the interest characteristics of the target users to obtain the target interest characteristics of the target platform users for the service subscription contents;

the content pushing based on the private domain flow pool of the target platform user according to the matched subscription content to be pushed comprises the following steps:

releasing the matched subscription content to be pushed through a preset interactive community of the target platform user, and setting the released subscription content to be pushed to be visible only to the interactive platform user in the private domain flow pool on the interactive community; or

Acquiring interaction platform users to be pushed, the interaction relationship of which with the target platform users reaches preset conditions, according to the interaction relationship description information among the user identification information, and pushing the subscription content to be pushed to the interaction platform users to be pushed through the target platform users;

the step of correcting the interestingness parameter included in the interest feature of the target user according to the interestingness parameter included in the interest feature of the interactive user of the service subscription content by each interactive platform user comprises the following steps:

determining an interaction heat coefficient between the target platform user and each interaction platform user according to the interaction relationship between the target platform user and each interaction platform user;

according to the interaction heat coefficient between the target platform user and each interaction platform user, carrying out weighted product on the interest parameters included by the interest characteristics of the interaction users of each interaction platform user aiming at the service subscription content, and then calculating an average value;

and correcting the interestingness parameter included in the interest feature of the target user in a mode of multiplying the calculated average value by a proportional coefficient of a preset threshold value and the interestingness parameter included in the interest feature of the target user.

In a second aspect, the embodiment of the present invention further provides a social network platform, which includes a processor, a machine-readable storage medium connected to the processor, the machine-readable storage medium being configured to store a program, an instruction, or code, and the processor being configured to execute the program, the instruction, or the code in the machine-readable storage medium to implement the above-mentioned method.

Compared with the prior art, the method and the device have the advantages that the interactive user interest characteristics of service subscription contents provided by the social network platform and the target user interest characteristics of the service subscription contents provided by the target platform user on the social network platform are obtained through analysis of the user subscription information knowledge base corresponding to the private domain flow pool of the target platform user to be subjected to content push, and then the content push is carried out on the service subscription contents provided by the social network platform based on the private domain flow pool of the target platform user according to the interactive user interest characteristics and the target user interest characteristics. Therefore, interest characteristic analysis of service subscription content is carried out on the target platform user and the interaction platform user having an effective interaction relation with the target platform user from different information dimensions, subscription content is pushed based on the result of the interest characteristic analysis, and therefore content pushing accuracy can be effectively improved, and content pushing effect is greatly improved.

In addition, because the private domain traffic pool is an effective private domain traffic generated by strict screening, for example, an effective interactive object having a large amount of interactive behaviors or a specific interactive record with a target platform user, when content is pushed based on the private domain traffic of the target platform user, the problem of potential traffic loss caused by user dislike due to improper content pushing can be avoided.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, it should be understood that the following drawings only illustrate some embodiments of the present invention, and therefore should not be considered as limiting the scope, and for those skilled in the art, other related drawings can be obtained according to the drawings without inventive efforts.

FIG. 1 is a schematic diagram of an operating environment of an embodiment of the present application.

Fig. 2 is a schematic flow chart of a content push method based on big data and private domain traffic according to an embodiment of the present application.

Fig. 3 is a flow chart illustrating the sub-steps of step S400 in fig. 2.

Fig. 4 is a schematic diagram of an apparatus for implementing the above method provided by an embodiment of the present application.

Fig. 5 is a functional module diagram of the content push device in fig. 4.

Detailed Description

The technical solutions in the present application will be described clearly and completely with reference to the accompanying drawings in the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

Referring to fig. 1, fig. 1 is a schematic diagram of an operating environment according to an embodiment of the present disclosure, where the operating environment relates to a social networking system 10, the social networking system 10 includes a social networking platform 100 and a plurality of terminal devices communicatively connected to the social networking platform 100, and the social networking platform 100 may also provide a client for social communication. In this way, the terminal device may perform social network communication with other terminal devices through a client provided by the social network platform 100, for example, sending of instant messages, group conversation, friend circle interaction, community interaction, forum interaction, and the like may be performed through the client. The social network platform 100 may be a server (e.g., a content push server), a server group, a distributed network system, a mainframe computer device, etc., and is not particularly limited herein. The client provided by the social network platform 100 may be, for example, an instant messaging client, a microblog client, a forum client, a public number client, and the like, which is not limited in this embodiment. The plurality of terminal devices may be, for example, the target terminal device 200 used by one target platform user and the other terminal devices 300 used by other platform users interacting with the presence information of the target platform user according to the embodiment of the present application. In the embodiment of the present application, the social network platform 100 may perform large data processing and analysis on a large amount of historical interaction information generated by information interaction between the target platform user and other platform users, so as to identify the effective private domain flow of the target platform user, and then establish a private domain flow pool for the target platform user, so as to perform works such as advertisement delivery and content push in a later period, so as to improve the flow operation effect based on the private domain flow pool and prevent potential flow loss and other problems. Further, the social network platform 100 is further configured to perform content push for a target platform user to perform content push according to the knowledge base corresponding to the established private domain traffic pool component. The detailed method is specifically exemplified below.

Referring to fig. 2, a schematic flow interaction diagram of a content pushing method based on big data and a private traffic pool according to an embodiment of the present application is provided, where the method is mainly described in detail below by the social network platform 100 in the social network system 10 executing specific steps included in the method.

Step S100, a private domain flow pool of a target platform user to be subjected to content push in the social network platform is obtained, and a corresponding user subscription information knowledge base is obtained according to the private domain flow pool.

In this embodiment, the user subscription information repository corresponding to the private domain traffic pool includes user identification information of an interaction platform user having an effective interaction relationship with the target platform user, content identification information of service subscription content provided by the social network platform, interaction relationship description information between the user identification information, and content subscription description information between the user identification information and the content identification information, where the interaction relationship description information includes interaction relationship information between the interaction platform users, and the content subscription description information includes historical content subscription information of the interaction platform user on the service subscription content provided by the social network platform. The interaction relationship description information may include, but is not limited to, a description of an interaction degree (e.g., an interaction score) between each interaction platform user and the target platform user, a description of an interaction degree (e.g., an interaction score) between each interaction platform user, interaction content between each platform user (including the interaction platform user and the target platform user), index information related to interaction time information, and the like, and is not limited specifically. The content subscription description information may also include, but is not limited to, subscription information related to each platform user for specific service subscription content, such as subscription operation information related to specific service subscription content, for example, subscription time, a subscription content tag, a type of subscription content, and the like.

Step S200, obtaining the interactive user interest characteristics of each interactive platform user in the private flow pool of the target platform user to the service subscription content provided by the social network platform according to the interactive relationship description information between the user identification information of the interactive platform user and the content subscription description information between the user identification information and the content identification information.

Step S300, obtaining the historical content subscription information of the target platform user, and obtaining the target user interest characteristics of the target platform user on the service subscription content provided by the social network platform according to the historical content subscription information.

And step S400, carrying out content push on the service subscription content provided by the social network platform based on the private domain flow pool of the target platform user according to the interactive user interest characteristics and the target user interest characteristics.

In summary, in the embodiments of the present invention, the interactive user interest characteristics of the service subscription content provided by the social network platform and the target user interest characteristics of the service subscription content provided by the target platform user on the social network platform are obtained through analyzing the user subscription information knowledge base corresponding to the private domain traffic pool of the target platform user to be content-pushed, and then the content-pushed is performed on the service subscription content provided by the social network platform based on the private domain traffic pool of the target platform user according to the interactive user interest characteristics and the target user interest characteristics. Therefore, interest characteristic analysis of service subscription content is carried out on the target platform user and the interaction platform user having an effective interaction relation with the target platform user from different information dimensions, subscription content is pushed based on the result of the interest characteristic analysis, and therefore content pushing accuracy can be effectively improved, and content pushing effect is greatly improved.

The relevant contents in the above steps will be further elaborated with reference to specific embodiments.

In the above step S100, the private traffic pool of the target platform user can be constructed through the following steps S1-S5, which are described in detail below.

Step S1, a first interactive object sequence recorded by a target platform user in the social network platform at a first time node is obtained.

In this embodiment, the first time node is any time node before the target platform user currently logs in the social network platform. The first interactive object sequence comprises two or more interactive platform users which have interactive relation with the target platform user. And each interactive platform user has a history information interaction record between the second time node and the target platform user. The second time node is a time node before the first time node. The first interactive object sequence is an interactive object sequence formed by a plurality of interactive platform users which are recorded by the target platform user at the second time node and interact with the target platform user existence information.

Step S2, sequentially using the two or more interactive platform users of the first interactive object sequence as platform users to be analyzed, and collecting historical interactive behavior information generated between each platform user to be analyzed and the target platform user within a target time interval.

Step S3, performing big data analysis based on the interaction behavior information, judging whether the interaction behavior between each platform user to be analyzed and the target platform user reaches a preset effective interaction condition, and if the interaction behavior reaches the preset effective interaction condition, determining that the platform user to be analyzed is an effective interaction platform user.

Step S4, marking the user identification information of the effective interactive platform user in the first interactive object sequence, and adding the user identification information of the effective interactive platform user to the effective interactive object sequence.

Step S5, adding each effective interaction platform user in the effective interaction object sequence as an effective private domain traffic of the target platform user to a private domain traffic pool corresponding to the target platform user in the social network platform.

In the sub-step S2, the interaction behavior information generated between the platform user to be analyzed and the target platform user may be collected in the target time interval by the following method, which is exemplarily described as follows:

firstly, determining the target time interval according to the first time node, wherein the target time interval comprises a preset time period before the target platform user logs in the social network platform and comprises the second time node;

then, acquiring historical interaction behavior information between the target platform user and each platform user to be analyzed based on historical interaction behaviors of the target platform user and each platform user to be analyzed in the target time interval;

and finally, using the collected historical interactive behavior information as the interactive behavior information generated between the platform user to be analyzed and the target platform user.

In this embodiment, the historical interaction behavior information is obtained by collecting interaction information based on the interaction behaviors of the target platform user and the platform user to be analyzed within the target time interval. The historical interaction behavior information may include at least one of the following three types of information: the analysis method comprises the steps of carrying out information interaction between a target platform user and a platform user to be analyzed to generate first type interaction information, carrying out information interaction between the target platform user and the platform user to be analyzed in a multi-person conversation group to generate second type interaction information, and sending third type interaction information to the platform user to be analyzed in an interaction community of the target platform user. Based on this, in step S3, performing big data analysis based on the interaction behavior information, and determining whether the interaction behavior between the platform user to be analyzed and the target platform user reaches a preset effective interaction condition, which may be implemented in the following manner:

and calculating to obtain interaction scores between the platform users to be analyzed and the target platform users according to the historical interaction behavior information, and judging whether the interaction behaviors between the platform users to be analyzed and the target platform users reach preset effective interaction conditions or not according to the interaction scores.

Further, in this embodiment, in a first possible implementation manner, the historical interaction behavior information may include first type interaction information generated by information interaction between the target platform user and the platform user to be analyzed. Based on this, in the step S3, an interaction score between each platform user to be analyzed and the target platform user is calculated according to the historical interaction behavior information, and whether the interaction behavior between each platform user to be analyzed and the target platform user reaches a preset effective interaction condition is determined according to the interaction score, and a specific implementation method may be implemented through the substeps S31-S34 shown in fig. 3, which is specifically described below.

And a substep S31, obtaining, for the first type of interaction information corresponding to each platform user to be analyzed, information interaction times between the platform user to be analyzed and the target platform user in the first type of interaction information. For example, the number of information interactions refers to the number or frequency of interaction information sent between the target platform user and each platform user to be analyzed. The first type of interaction information may be, for example, interaction information generated by the target platform interacting with the platform to be analyzed through an Instant Messaging (IMS) provided by the client of the social network platform 100, such as private chat information including text, pictures, and voice.

And a substep S32, matching the first type of interaction information with a pre-established keyword library, and extracting target keywords stored in the keyword library contained in the first type of interaction information. In this embodiment, the keyword library may include a large number of pre-collected keywords that may be used to represent the degree of closeness between different platform users, and may be used to assist in identifying the effective private domain traffic of the target platform user. For example, the keyword library may include "lovely", "brother", "dad", "mom", "old classmate", "brother", "old iron", and the like.

And a substep S33, performing interactive information quantity feature extraction on the first type of interactive information to obtain interactive information quantity features corresponding to the first type of interactive information. In this embodiment, the interactive information amount characteristic may include an information size (for example, expressed in bytes) and a text length of the first type of interactive information. The interactive information quantity characteristic is used for representing the information quantity of information interaction between the target platform user and the platform user to be analyzed, and the larger the information quantity is, the more the interaction between the target platform user and the platform user to be analyzed is, otherwise, the less the interaction is. Therefore, the interactive information quantity characteristics can also be used for assisting in identifying the effective private domain flow of the target platform user, and the identification accuracy of the effective private domain flow is improved.

And a substep S34, calculating the information interaction times, the target keywords and the interaction information amount score according to the first type of interaction information corresponding to each platform user to be analyzed to obtain an interaction score between each platform user to be analyzed and the target platform user, and judging whether the interaction behavior between each platform user to be analyzed and the target platform user reaches a preset effective interaction condition according to the interaction score.

In another possible implementation manner, the historical interaction behavior information may further include first type interaction information generated by information interaction between the target platform user and the platform user to be analyzed, and second type interaction information generated by information interaction between the target platform user and the platform user to be analyzed in a multi-person conversation group. Based on this, in the step S3, an interaction score between each platform user to be analyzed and the target platform user is calculated according to the historical interaction behavior information, and whether the interaction behavior between each platform user to be analyzed and the target platform user reaches a preset effective interaction condition is determined according to the interaction score, and a specific implementation manner may include the following steps (1) to (4), which are described in detail below.

(1) And calculating to obtain a first interaction behavior score according to the first type interaction information aiming at the historical interaction behavior information corresponding to each platform user to be analyzed.

In this embodiment, a first interaction behavior score is obtained by calculating the first type of interaction information, and an implementation manner is described as follows:

firstly, acquiring the information interaction times between the platform user to be analyzed and the target platform user in the first type of interaction information, and calculating according to the information interaction times and a preset integral rule to obtain an interaction time score;

then, matching the first type of interaction information with a pre-established keyword library, extracting target keywords stored in the keyword library contained in the first type of interaction information, and matching the extracted target keywords with a preset keyword integral mapping table to obtain keyword scores corresponding to the first type of interaction information;

then, extracting interactive information quantity characteristics of the first type of interactive information to obtain interactive information quantity characteristics corresponding to the first type of interactive information, and calculating to obtain an interactive information quantity score according to the interactive information quantity characteristics corresponding to the first type of interactive information;

and finally, calculating to obtain a first interaction behavior score between each platform user to be analyzed and the target platform user according to the interaction times score, the keyword score and the interaction information amount score.

(2) And calculating to obtain a second interaction behavior score according to the second type interaction information.

In this embodiment, the second interaction behavior score is obtained by calculation according to the second type of interaction information, and an implementation manner is described as follows:

firstly, acquiring the number of shared groups of a multi-person conversation group shared by each platform user to be analyzed and the target platform user respectively, the number of times of attention paid to each other by each platform user to be analyzed and the target platform user in the multi-person conversation group, and the interactive information quantity of the interactive information sent by each platform user to be analyzed and the target platform user in the shared multi-person conversation group from the second type interactive information; the operation of paying attention to each other may be, for example, "@ opponent", "one beat opponent", or the like;

and then, calculating to obtain a second interaction behavior score between each platform user to be analyzed and the target platform user according to the number of the common groups, the attention times and the interaction information amount.

(3) And calculating to obtain the interaction scores of the platform users to be analyzed and the target platform users according to the first interaction behavior score and the second interaction behavior score.

(4) And judging whether the interaction behavior between each platform user to be analyzed and the target platform user reaches a preset effective interaction condition or not according to the interaction score between each platform user to be analyzed and the target platform user.

Further, in another implementation of this embodiment, the historical interaction behavior information may further include first type interaction information generated by information interaction between the target platform user and the platform user to be analyzed, second type interaction information generated by information interaction between the target platform user and the platform user to be analyzed in a multi-person conversation group, and third type interaction information sent by the platform user to be analyzed in an interaction community of the target platform user. The interactive community may be, for example, an interactive space established by the target platform user using a client provided by the social network platform 100, such as, but not limited to, a circle of friends, a microblog comment area, a forum, a post, etc. Based on this, in the above sub-step S3, an interaction score between each platform user to be analyzed and the target platform user is calculated according to the historical interaction behavior information, and whether the interaction behavior between each platform user to be analyzed and the target platform user reaches a preset effective interaction condition is determined according to the interaction score, and a specific implementation method may be implemented through the following steps (1) - (5).

In this implementation, the calculating a first interaction behavior score according to the first type of interaction information includes:

acquiring the information interaction times between the platform user to be analyzed and the target platform user in the first type of interaction information, and calculating according to the information interaction times and a preset integral rule to obtain an interaction time score;

matching the first type interaction information with a pre-established keyword library, extracting target keywords stored in the keyword library contained in the first type interaction information, and matching the extracted target keywords with a preset keyword integral mapping table to obtain keyword scores corresponding to the first type interaction information;

extracting interactive information quantity characteristics of the first type of interactive information to obtain interactive information quantity characteristics corresponding to the first type of interactive information, and calculating to obtain an interactive information quantity score according to the interactive information quantity characteristics corresponding to the first type of interactive information;

and calculating to obtain a first interaction behavior score between each platform user to be analyzed and the target platform user according to the interaction times score, the keyword score and the interaction information amount score.

In this embodiment, the calculating a second interaction behavior score according to the second type of interaction information includes:

acquiring the number of shared groups of a multi-person conversation group shared by each platform user to be analyzed and the target platform user respectively, the number of times of attention paid to each other by each platform user to be analyzed and the target platform user in the multi-person conversation group, and the interactive information amount of the interactive information sent by each platform user to be analyzed and the target platform user in the shared multi-person conversation group from the second type interactive information;

and calculating to obtain a second interaction behavior score between each platform user to be analyzed and the target platform user according to the number of the common groups, the attention times and the interaction information amount.

(3) And calculating to obtain a third interaction behavior score according to the third type interaction information.

In this embodiment, the calculating a third interaction behavior score according to the third type of interaction information includes:

acquiring target operation frequency with a target operation label executed by each platform user to be analyzed in an interactive community of the target platform user, interactive information quantity of interactive information sent in an information interactive area of the interactive community and target keywords included in the interactive information from the third type of interactive information; the target operation may be, for example, approval, sending of a virtual gift, paying attention to each other, etc.; the interactive information may be, for example, interactive information transmitted in the interactive area;

and calculating to obtain a third interactive behavior score between each platform user to be analyzed and the target platform user according to the target operation frequency, the interactive information amount and the target keywords.

(4) And calculating to obtain the interaction scores of the platform users to be analyzed and the target platform users according to the first interaction behavior score, the second interaction behavior score and the third interaction behavior score.

(5) And judging whether the interaction behavior between each platform user to be analyzed and the target platform user reaches a preset effective interaction condition or not according to the interaction score between each platform user to be analyzed and the target platform user.

In this embodiment, in each of the above methods, whether the interaction behavior between each platform user to be analyzed and the target platform user reaches a preset effective interaction condition may be determined according to the interaction score between each platform user to be analyzed and the target platform user in any one of the following two ways, which are specifically described below.

According to the first scheme, each platform user to be analyzed is sorted in a descending order according to the interaction score between each platform user to be analyzed and the target platform user, and a preset number of platform users to be analyzed, which are sorted in the front, are determined as interaction behaviors with the target platform user to reach a preset effective interaction condition according to a sorting result.

And judging whether the interaction scores between each platform user to be analyzed and the target platform user respectively reach a preset score threshold value, if so, judging that the interaction behavior between the corresponding platform user to be analyzed and the target platform user reaches a preset effective interaction condition, and if not, judging that the interaction behavior between the corresponding platform user to be analyzed and the target platform user does not reach the preset effective interaction condition.

Therefore, the embodiment of the invention can further improve the identification accuracy of the private domain flow of the target platform user by finely analyzing the effective interaction behavior between the platform user to be analyzed and the target platform user through different types of historical interaction information.

Further, in step S100, constructing the user subscription information repository according to the private domain traffic pool may be implemented by the following method, which is described in detail as follows.

Firstly, the interaction relationship between different interaction platform users in the private flow pool and the target platform user and the interaction relationship between the interaction platform users can be obtained, and a platform user interaction relationship matrix is obtained;

then, obtaining historical subscription information of each target platform user and each interaction platform user on subscription content provided by the social network platform from the social network platform to obtain a historical subscription information matrix;

and finally, constructing the user subscription information knowledge base according to the platform user interaction relation matrix and the historical subscription information matrix, wherein the user subscription information knowledge base can be a knowledge graph. For example, each platform user may be used as a node in a knowledge graph according to the platform user interaction relationship matrix, platform users are connected according to the interaction relationship, and then each historical subscription information is used as an information node, and the connection is performed according to the subscription relationship between each platform user and the historical subscription information, so as to form the user subscription information knowledge base. In the user subscription information knowledge base, the connection line between the platform users and the subscription information is subscription information (for example, the subscription frequency of specific subscription content, etc.), and the interaction relationship (for example, the interaction score) between the platform users corresponding to the connection line between the platform users is obtained.

Further, in the step S200, a manner of obtaining the interactive user interest characteristics of the service subscription content provided by each interactive platform user in the private domain traffic pool of the target platform user to the social network platform may be implemented in the following manner.

Firstly, converting interactive relationship description information among user identification information of each interactive platform user into an interactive relationship characteristic matrix through a knowledge base deep learning network obtained through pre-training, and converting content subscription description information between the user identification information and content identification information into a content subscription characteristic matrix.

And then, obtaining the interactive user interest characteristics of each interactive platform user in the private domain flow pool of the target platform user on the service subscription content provided by the social network platform according to the interactive relationship characteristic matrix and the content subscription characteristic matrix.

In detail, first, a preset traversal node combination sequence for performing information traversal in a user subscription information knowledge base corresponding to the private domain traffic pool may be obtained, where the traversal node combination sequence includes a plurality of different traversal node combinations. For example, all node paths in the knowledge base including the root node may be sequentially traversed starting from the root node of the knowledge base, and the nodes included in each node path may be combined as one traversal node.

Then, according to all traversal node combinations in the traversal node combination sequence, information extraction is sequentially carried out in a user subscription information knowledge base corresponding to the private domain flow pool, and a plurality of user identification information sets corresponding to the traversal node combinations respectively and a plurality of content subscription description information sets corresponding to the traversal node combinations respectively are obtained;

and finally, the plurality of user identification information sets and the plurality of content subscription description information sets can be respectively input into the knowledge base deep learning network for information processing, so that the interaction relationship characteristic matrix and the content subscription characteristic matrix are obtained.

In addition, in this embodiment, the knowledge base deep learning network may be obtained by training in advance in the following manner, which is described in detail below.

Firstly, acquiring a knowledge base sample, wherein the knowledge base sample comprises sample user identification information of sample interaction platform users, sample content identification information of service subscription content provided by a sample social network platform, interaction relation description information between the sample user identification information and content subscription description information between the sample user identification information and the sample content identification information, the interaction relation description information comprises interaction relation information between the sample interaction platform users, and the content subscription description information comprises historical content subscription information of the sample interaction platform users on the service subscription content provided by the sample social network platform.

And then, converting the identification information of each sample user into a sample interaction relation characteristic matrix through the knowledge base deep learning network, and converting content subscription description information between the identification information of the sample user and the identification information of the sample content into a sample content subscription characteristic matrix.

And finally, performing iterative training on the knowledge base deep learning network according to each sample interaction relation characteristic matrix and each sample content subscription characteristic matrix until the network model loss function value of the knowledge base deep learning network obtained in the iterative training process reaches a preset training termination condition.

Further, in step S300, the target user interest characteristics of the target platform user in the service subscription content provided by the social network platform may be obtained by: firstly, according to the subscription time of the target platform user for each piece of historical content subscription information, performing information processing on each piece of historical content subscription information to obtain a time sequence-based historical subscription characteristic matrix of the target platform user; and then, acquiring target user interest characteristics of the target platform user on service subscription content provided by the social network platform according to the historical subscription characteristic matrix and the content subscription characteristic matrix. For example, the history subscription feature matrix and the content subscription feature matrix may be subjected to fusion analysis processing according to a time sequence node, and a subscription content tag corresponding to an intersection of the history subscription feature matrix and the content subscription feature matrix is used as the target user interest feature.

Further, as shown in fig. 3, in the step S400, according to the interactive user interest feature and the target user interest feature, content pushing is performed on the service subscription content provided by the social network platform based on the private traffic pool of the target platform user, and a specific implementation method may include the sub-steps S401 to S403 shown in fig. 3, which are described in detail as follows.

And a substep S401 of dividing a plurality of interactive platform users, the interaction relationships of which meet preset conditions, into an interest group to be pushed according to the interaction relationship description information among the user identification information, and taking the content tags of the historical subscription content of each platform user in the interest group as the interest characteristics of the interactive users corresponding to the interest group. For example, the interaction platform users with interaction scores of 80-100 points can be divided into the interest groups according to the interaction scores of the interaction platform users with the target platform users.

And a substep S402, according to the interest characteristics of each interactive user, matching corresponding subscription content to be pushed from the social network platform, and pushing the subscription content to be pushed to the interactive platform users in the interest group corresponding to the interest characteristics of each interactive user through a first content pushing rule predetermined based on the target platform user. For example, similar subscribed content which is not subscribed can be searched as the subscribed content to be pushed according to the interactive user interest characteristics.

Further, an alternative implementation of content push according to the first content push rule may be as follows: and sending corresponding subscription content to be pushed to the social network clients of the interaction platform users in the interest group corresponding to the interest characteristics of the interaction users through the social network clients of the target platform users. For example, on the premise of obtaining the consent of the target platform user in advance, the content to be pushed and subscribed is directly sent to the social network client of the interaction platform user in the interest group through the social network client of the target platform user. For example, in actual implementation, a content push request may be first sent to the target platform user, where the content push request includes specific subscription content to be pushed and an interaction platform user to which the subscription content to be pushed is targeted, and then when feedback of the target platform user through the content push request is received, the social network platform directly starts, in the background, to send push information of the subscription content to be pushed to the social network clients of the interaction platform users in the interest group through the social network clients of the target platform user. Therefore, the interactive platform users in the interest group are users in the effective private flow pool of the target platform user, and the interaction score with the target platform user is higher, so that the users can be approved by the push platform user in a mode of directly sending push content by the target platform user, the content push accuracy is higher, and the push effect is better.

And a substep S403, matching corresponding subscription content to be pushed from the social network platform according to the target user interest characteristics, and pushing the subscription content to be pushed to at least one interactive platform user in the private domain traffic pool through a second content pushing rule predetermined based on the target platform user.

In this embodiment, similar subscription content that is not subscribed by each interactive platform user may also be searched as subscription content to be pushed according to the interactive user interest characteristics.

Based on this, in sub-step S403, the subscription content to be pushed is pushed to at least one interactive platform user in the private domain traffic pool according to a second content pushing rule predetermined based on the target platform user, and a specific implementation manner may include any one of the following two manners.

In the first mode, the subscription content to be pushed is released through a preset interactive community of the target platform user, and the released subscription content to be pushed is set to be visible only to the interactive platform user in the private domain flow pool on the interactive community. For example, the subscription content to be pushed can be directly published on an interaction community such as a friend group, a community platform, and a forum provided by the social network client of the target platform user, and the published content is set to be visible only to the user in the private domain traffic pool during publication, so that interference with some invalid private domain traffic can be avoided, and accurate pushing of the content can be realized. The push mode can be understood as that content is pushed to each interactive platform user in the private domain traffic pool.

And in the second mode, according to the interactive relationship description information among the user identification information, acquiring the interactive platform user to be pushed, the interactive relationship of which with the target platform user reaches the preset condition, and pushing the subscription content to be pushed to the interactive platform user to be pushed through the target platform user. For example, the interaction platform user whose interaction score between the target platform users reaches a preset score threshold may be determined as the interaction platform user to be pushed, and then the subscription content to be pushed is directly published on the interaction communities such as the friend circle, the community platform, and the forum provided by the social network client of the target platform user, and the published content is set to be visible only by the interaction platform user to be pushed when the subscription content is published, so that the content is pushed accurately. The push mode can be understood as content push to a part of accurate interactive platform users in the private domain traffic pool. In another implementation manner, the subscription content to be pushed may also be directly sent to the social network client of the interactive platform user to be pushed through the social network client of the target platform user.

In another alternative embodiment, in step S400, according to the interactive user interest feature and the target user interest feature, content push is performed on the service subscription content provided by the social network platform based on the private traffic pool of the target platform user, which may also be implemented through the following steps (1) - (4), which are described in detail below.

(1) And performing feature fusion on the interactive user interest features of each interactive platform user in the private flow pool of the target platform user on each service subscription content provided by the social network platform and the target user interest features of the target platform user for each service subscription content provided by the social network platform to obtain the target interest features of the target platform user on the service subscription content provided by each social network platform, wherein the target interest features comprise interest degree parameters of the target platform user for each service subscription content. In detail, for each service subscription content, a target user interest characteristic of the target platform user for the service subscription content and an interactive user interest characteristic of each interactive platform user for the service subscription content may be obtained; and then, according to the interest degree parameters included in the interest characteristics of the interaction users of the service subscription content of each interaction platform user, correcting the interest degree parameters included in the interest characteristics of the target users to obtain the target interest characteristics of the target platform users for the service subscription content.

For example, the calibration method may be: and when the average value of the interest degree parameters included in the interest characteristics of the interactive users of the service subscription content by each interactive platform user is greater than a preset threshold value, promoting the interest degree parameters included in the interest characteristics of the target users, for example, promoting according to a preset promoting rule. For example, the target user interest feature may be increased by a ratio exceeding a preset threshold, for example, by 20%, and then the interestingness parameter included in the target user interest feature is correspondingly increased by 20%.

For another example, the interactive heat coefficient between the target platform user and each interactive platform user may be determined according to the interactive relationship between the target platform user and each interactive platform user. For example, the interaction heat coefficient corresponding to the interaction platform user with the largest interaction score is 1, and the interaction heat coefficients corresponding to other interaction platform users can be calculated according to the proportion. For example, the maximum interaction score of the interaction platform user a and the target platform user is 100 points, the interaction heat coefficient is 1, the score of the interaction platform user B and the target platform user is 75 points, the interaction heat coefficient corresponding to the interaction platform user B is 75/100 and is equal to 0.75 point, and the other steps are calculated accordingly.

And then, according to the interactive heat coefficient between the target platform user and each interactive platform user, carrying out weighted multiplication on the interest degree parameters included by the interest characteristics of the interactive users of each interactive platform user aiming at the service subscription content, and then obtaining an average value.

And finally, correcting the interestingness parameter included in the interest feature of the target user in a mode of multiplying the calculated average value by a proportional coefficient of a preset threshold value and the interestingness parameter included in the interest feature of the target user.

Therefore, content pushing is performed subsequently based on the corrected interestingness parameter, and the content pushing accuracy can be higher.

(2) And according to the target interest characteristics of the service subscription content provided by the target platform user to each social network platform, performing descending arrangement according to the interest degree parameters.

(3) And selecting a preset number of target interest characteristics to match the subscription content to be pushed in the social network platform according to the target interest characteristics after the descending order arrangement. For example, a preset number of top-ranked target interest features may be selected to match the subscription content to be pushed on the social network platform.

(4) And pushing the content based on the private domain flow pool of the target platform user according to the matched subscription content to be pushed. For example, the matched subscription content to be pushed can be released through a preset interactive community of the target platform user, and the released subscription content to be pushed is set to be visible only to the interactive platform user in the private traffic pool on the interactive community; or acquiring the interactive platform users to be pushed, the interactive relationship of which with the target platform users reaches the preset condition, according to the interactive relationship description information among the user identification information, and pushing the subscription content to be pushed to the interactive platform users to be pushed through the target platform users.

Fig. 4 is a schematic view of a device for implementing the content push method based on big data and private domain traffic according to an embodiment of the present invention, where the device may be the social network platform 100 or a content push server in the social network system 10, and preferably, in this embodiment, the device is the social network platform 100. The social network platform 100 may be, but is not limited to, a server, a personal computer, a server cluster, and the like with big data analysis and processing capability.

The apparatus may include a content push device 110, a machine-readable storage medium 120, and a processor 130.

In this embodiment, the machine-readable storage medium 120 and the processor 130 may be communicatively coupled to each other and accessed by the processor 130 via a bus interface. Alternatively, the machine-readable storage medium 120 may be integrated into the processor 130, e.g., may be a cache and/or general purpose registers.

The content pushing device 110 may include software functional modules (e.g., the software functional modules included in the content pushing device 110) stored in the machine-readable storage medium 120. The method provided by the foregoing method embodiment may be implemented when the processor 130 executes a software function module in the content pushing device 110.

In detail, as shown in fig. 5, the content pushing apparatus 110 includes a plurality of software functional modules, such as a knowledge base obtaining module 111, a first interest feature analysis module 112, a second interest feature analysis module 113, and a subscribed content pushing module 114.

The knowledge base obtaining module 111 is configured to obtain a private domain traffic pool of a target platform user to be content-pushed in the social network platform, and obtain a corresponding user subscription information knowledge base according to the private domain traffic pool.

A first interest feature analysis module 112, configured to obtain, according to the interaction relationship description information between the user identification information of the interaction platform users and the content subscription description information between the user identification information and the content identification information, an interest feature of each interaction platform user in the private domain traffic pool of the target platform user for an interaction user of the service subscription content provided by the social network platform.

The second interest feature analysis module 113 is configured to obtain historical content subscription information of the target platform user, and obtain a target user interest feature of the service subscription content provided by the target platform user to the social network platform according to the historical content subscription information.

And a subscription content pushing module 114, configured to, according to the interactive user interest feature and the target user interest feature, perform content pushing on service subscription content provided by the social network platform based on a private domain traffic pool of the target platform user.

It should be noted that the content pushing apparatus 110 includes a plurality of software functional modules, such as a knowledge base obtaining module 111, a first interest feature analyzing module 112, a second interest feature analyzing module 113, and a subscribed content pushing module 114, which can be respectively used to implement the steps S100, S200, S300, S400, and the like shown in fig. 2. For specific implementation methods and more detailed contents of the functional modules, reference may be made to the specific contents of steps S100, S200, S300, and S400, which are not described herein again.

The embodiments described above are only a part of the embodiments of the present invention, and not all of them. The components of embodiments of the present invention generally described and illustrated in the figures can be arranged and designed in a wide variety of different configurations. Therefore, the detailed description of the embodiments of the present invention provided in the drawings is not intended to limit the scope of the present invention, but is merely representative of selected embodiments of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims. Moreover, all other embodiments that can be made available by a person skilled in the art without inventive step based on the embodiments of the present invention shall fall within the scope of protection of the present invention.

Claims

1. A content pushing method based on big data and private domain flow is characterized by comprising the following steps:

2. The method of claim 1, wherein the content pushing of the service subscription content provided by the social networking platform based on the private traffic pool of the target platform user according to the interactive user interest feature and the target user interest feature comprises:

3. The method according to claim 2, wherein the pushing the subscription content to be pushed to the interactive platform users in the interest group corresponding to the interest feature of each interactive user through a first content pushing rule predetermined based on the target platform user comprises:

4. The method of claim 1, further comprising:

5. The method according to claim 1, wherein the obtaining of the interactive user interest characteristics of each interactive platform user in the private traffic pool of the target platform user for the service subscription content provided by the social network platform according to the interaction relationship description information between the user identification information of the interactive platform user and the content subscription description information between the user identification information and the content identification information comprises:

6. The method of claim 5, wherein the obtaining of the target user interest characteristics of the target platform user for the service subscription content provided by the social network platform according to the historical subscription characteristic matrix and the content subscription characteristic matrix comprises:

7. The method of claim 5, further comprising:

8. The method of claim 5, further comprising:

9. The method of claim 1, wherein the pushing content of the service subscription provided by the social networking platform based on the private traffic pool of the target platform user according to the interactive user interest feature and the target user interest feature comprises:

10. A social networking platform comprising a processor, a machine-readable storage medium coupled to the processor, the machine-readable storage medium configured to store a program, instructions, or code, the processor configured to execute the program, instructions, or code in the machine-readable storage medium to implement the method of any one of claims 1-9.