CN113268589B - Key user identification method, key user identification device, readable storage medium and computer equipment - Google Patents

Key user identification method, key user identification device, readable storage medium and computer equipment Download PDF

Info

Publication number
CN113268589B
CN113268589B CN202010092848.7A CN202010092848A CN113268589B CN 113268589 B CN113268589 B CN 113268589B CN 202010092848 A CN202010092848 A CN 202010092848A CN 113268589 B CN113268589 B CN 113268589B
Authority
CN
China
Prior art keywords
game
dimension
user
game player
features
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010092848.7A
Other languages
Chinese (zh)
Other versions
CN113268589A (en
Inventor
金诚
刘筱叶
蔡红云
何峰
姚亮
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tencent Technology Shenzhen Co Ltd
Original Assignee
Tencent Technology Shenzhen Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Tencent Technology Shenzhen Co Ltd filed Critical Tencent Technology Shenzhen Co Ltd
Priority to CN202010092848.7A priority Critical patent/CN113268589B/en
Publication of CN113268589A publication Critical patent/CN113268589A/en
Application granted granted Critical
Publication of CN113268589B publication Critical patent/CN113268589B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/35—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
    • G06F16/36—Creation of semantic tools, e.g. ontology or thesauri
    • G06F16/367—Ontology
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Animal Behavior & Ethology (AREA)
  • Computational Linguistics (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The application relates to a key user identification method, a key user identification device, a computer readable storage medium and computer equipment, wherein the characteristics of the active dimension and the social dimension of a game player are extracted according to the active data and the social data of the game player in a database; obtaining game characteristics of each game according to the game knowledge graph; performing feature cross combination on the features of the active dimension of the game player, the features of the social dimension and the game features of each game to obtain combined features of the game player after multi-dimensional cross; performing cluster analysis according to the combination characteristics of the game players after multi-dimensional intersection to obtain user images of the game players in preset dimensions; according to different games, corresponding identification conditions are determined, and key users corresponding to the games are identified from the game players according to the attention dimension of the identification conditions and the user portraits of the game players in preset dimensions, so that the accuracy of identifying the key users for different games is improved.

Description

Key user identification method, key user identification device, readable storage medium and computer equipment
Technical Field
The present application relates to the field of computer technologies, and in particular, to a method and apparatus for identifying a key user, a computer readable storage medium, and a computer device.
Background
The advertisement message may be pushed to the relevant player in the initial stage of the new game being online or when the old game wants to find a new user. The new game hopes to acquire the original, early, influential game player; the old game wants to acquire potential influential gamers, or acquire influential gamers in similar games, which are key users in the pull-up phase of the game, to decide whether to acquire enough gamers or to let the gamers invite more gamers to join in one of the key factors. However, due to limited advertising resources, how to accurately find key users of games, especially key users loving sharing, is a core problem of game pull-up.
At present, game subsystems are divided for different games based on the drawing of a game knowledge graph, so that experts and users can conduct questionnaire investigation and scoring on the game subsystems, hundreds of dimensions of game features are designed, and each game has one game feature which can be represented; further, the players play games, the game features are weighted according to the game time length of the players, the game features are obtained through weighted summation, finally, a training set and a prediction set of machine learning are built for the player features and the game features, and the machine learning method is adopted for training; and predicting whether the player will go to play for the new game to obtain the corresponding key user for pulling new. The game has different attributes, such as a national game, a public game and the like, and the accuracy is lower when key users are identified for the games with different attributes.
Disclosure of Invention
Based on this, it is necessary to provide a key user identification method, apparatus, computer-readable storage medium and computer device for the problem of low accuracy when key users are identified for games of different attributes.
A method of key user identification, comprising:
extracting the characteristics of the active dimension of the game player according to the active data of the game player in the database;
extracting the characteristics of the social dimension of the game player according to the social data of the game player in the database;
extracting features of each game according to the game knowledge graph to obtain game features of each game;
performing feature cross combination according to the features of the active dimension of the game player, the features of the social dimension and the game features of each game to obtain combined features of the game player after multi-dimensional cross;
performing cluster analysis according to the active dimension characteristics, the social dimension characteristics and the multi-dimension crossed combined characteristics of the game player to obtain a user image of the game player in a preset dimension;
and identifying the corresponding key user from the game player based on the user image of the game player in the preset dimension according to the received attention dimension of the identification condition.
In one embodiment, the step of identifying the corresponding key user from the game player based on the user image of the game player in the preset dimension according to the received attention dimension of the identification condition includes:
according to the received identification conditions, determining the attention dimension of the identified key user;
acquiring a user portrait of the game player in a concerned dimension;
and filtering the game player according to the received identification conditions and the user image of the game player in the attention dimension to obtain a key user.
In one embodiment, the step of filtering the game player according to the received identification condition and the user image of the game player in the attention dimension to obtain the key user includes:
filtering the game player according to the received identification conditions and the user images of the game player in the concerned dimension, and filtering the game player to obtain a preselected user;
and analyzing the probability of playing the target game by the preselected user according to the user portrait of the preselected user in the preset dimension, and determining the user with the probability of playing the target game larger than a preset value in the preselected user as a key user.
In one embodiment, the step of identifying the corresponding key user from the game player based on the user image of the game player in the preset dimension according to the received attention dimension of the identification condition includes:
when a seed user is provided in the received identification condition, acquiring a user portrait of the seed user in a preset dimension;
according to the user portrait of the seed user in the preset dimension, determining the concerned dimension and the user portrait corresponding to the concerned dimension;
and determining users with similar user portraits to the seed users from the game players according to the attention dimension and the user portraits corresponding to the attention dimension as key users.
In one embodiment, the step of obtaining the user representation of the game player in the preset dimension by performing cluster analysis according to the active dimension feature, the social dimension feature and the multi-dimension cross combined feature of the game player includes:
respectively carrying out cluster analysis according to the active dimension characteristics, the social dimension characteristics and the multi-dimension crossed combined characteristics of the game player to obtain user figures of the game player in each dimension;
According to the characteristics of the active dimension of the game player, obtaining the game characteristics of the game logged in by the game player;
and clustering according to game features of the games played by the game players to obtain game preference portraits of the game players.
In one embodiment, the method further comprises:
processing the characteristics of the social dimension of the game player by adopting a graph embedding method and a graph representation learning method to obtain network embedded characteristics of the game player;
network embedded features of the gamer are added as features of the gamer's social dimension.
In one embodiment, the step of extracting features of each game according to the game knowledge graph to obtain game features of each game includes any one or more of the following steps:
first kind: constructing n-dimensional vectors of each game according to the dimensions of the subsystems, and acquiring game features of each game;
second kind: according to the game knowledge graph of each game, representing each game into a vector embedded based on a network, and obtaining the game characteristics of each game;
third kind: and finding and arranging similar games of the current game through a related similarity calculation method, and obtaining game characteristics of each game.
A key user identification device comprising:
the active feature extraction module is used for extracting the features of the active dimension of the game player according to the active data of the game player in the database;
the feature extraction module of the social dimension is used for extracting the features of the social dimension of the game player according to the social data of the game player in the database;
the game feature extraction module is used for extracting features of each game according to the game knowledge graph to obtain game features of each game;
the feature cross combination module is used for carrying out feature cross combination according to the features of the active dimension, the features of the social dimension and the game features of each game of the game player to obtain combined features of the game player after multi-dimension cross;
the clustering module is used for carrying out clustering analysis according to the active dimension characteristics, the social dimension characteristics and the multi-dimension crossed combined characteristics of the game player to obtain user images of the game player in the preset dimension;
and the key user identification module is used for identifying the corresponding key user from the game player based on the user image of the game player in the preset dimension according to the received attention dimension of the identification condition.
A computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of the method.
A computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of the method.
The key user identification method, the key user identification device, the computer readable storage medium and the computer equipment extract the characteristics of the active dimension and the social dimension of the game player according to the active data and the social data of the game player in the database; obtaining game characteristics of each game according to the game knowledge graph; performing feature cross combination on the features of the active dimension of the game player, the features of the social dimension and the game features of each game to obtain combined features of the game player after multi-dimensional cross; performing cluster analysis according to the combination characteristics of the game players after multi-dimensional intersection to obtain user images of the game players in preset dimensions; according to different games, corresponding identification conditions are determined, and key users corresponding to the games are identified from the game players according to the attention dimension of the identification conditions and the user portraits of the game players in preset dimensions, so that the accuracy of identifying the key users for different games is improved.
Drawings
FIG. 1 is a diagram of an application environment for a key user identification method in one embodiment;
FIG. 2 is a flow diagram of a method of key user identification in one embodiment;
FIG. 3 is a flow chart illustrating one of the steps of a key user identification method in one embodiment;
FIG. 4 is a schematic diagram of a data integration architecture of a key user identification method in one embodiment;
FIG. 5 is a block diagram of a key user identification device in one embodiment;
FIG. 6 is a block diagram of a key subscriber identity device according to another embodiment;
FIG. 7 is a block diagram of a computer device in one embodiment.
Detailed Description
The present application will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present application more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the application.
FIG. 1 is a diagram of an application environment for a key user identification method in one embodiment. Referring to fig. 1, the key user identification method is applied to a server 120. The terminal 110 and the server 120 are connected through a network. When the key user identification method is applied to the terminal 110, the terminal 110 extracts the characteristics of the active dimension of the game player according to the active data of the game player in the database; extracting features of social dimensions of the game player according to social data of the game player in the database; extracting features of each game according to the game knowledge graph to obtain game features of each game; performing feature cross combination on the features of the active dimension of the game player, the features of the social dimension and the game features of each game to obtain combined features of the game player after multi-dimensional cross; performing cluster analysis according to the active dimension characteristics, the social dimension characteristics and the multi-dimension crossed combined characteristics of the game player to obtain user images of the game player in the preset dimension; the terminal 110 identifies the corresponding key user from the game player based on the user portraits of the game player in the preset dimension according to the received attention dimension of the identification condition.
When the key user identification method is applied to the server 120, the server 120 extracts the features of the active dimension of the game player according to the active data of the game player in the database; extracting features of social dimensions of the game player according to social data of the game player in the database; extracting features of each game according to the game knowledge graph to obtain game features of each game; performing feature cross combination on the features of the active dimension of the game player, the features of the social dimension and the game features of each game to obtain combined features of the game player after multi-dimensional cross; performing cluster analysis according to the active dimension characteristics, the social dimension characteristics and the multi-dimension crossed combined characteristics of the game player to obtain user images of the game player in the preset dimension; the server 120 receives the recognition condition transmitted from the terminal 110, and recognizes the corresponding key user from the game player based on the user representation of the game player in the preset dimension according to the attention dimension of the received recognition condition.
The terminal 110 may be a desktop terminal or a mobile terminal, and the mobile terminal may be at least one of a mobile phone, a tablet computer, a notebook computer, and the like. The server 120 may be implemented as a stand-alone server or as a server cluster composed of a plurality of servers.
As shown in fig. 2, in one embodiment, a key user identification method is provided. The present embodiment is mainly exemplified by the application of the method to the terminal 110 in fig. 1. Referring to fig. 2, the key user identification method specifically includes the following steps:
step S220, extracting the characteristics of the active dimension of the game player according to the active data of the game player in the database.
Wherein the database stores active data, social data and the like generated by all game players in each game. The game player may be a user who logs in to any game or games in a game pad, which is the sum of all games. The active data of the game player refers to game information, login days information and login time information which the game player logs in. The feature of the active dimension of the game player is obtained by processing and integrating the game information, login days information and login duration information which are logged in by the game player.
The method comprises the steps of obtaining game information, login day information and login duration information which are logged in by a game player in a preset time period from a database, wherein the preset time period can be determined according to actual requirements, can be within about 30 days, within about 7 days, within about 14 days and the like, the game information can be identification information such as names of games, the login day information is login records generated when the game player logs in each game, and the login duration information is online duration records after the game player logs in each game. Processing the game information, the login days information and the login time information logged in by the game player, such as: determining the name of a game played by a game player according to the game information logged in by the game player; removing repeated login time in login records generated when a game player logs in each game, and calculating and obtaining total days of playing the game in a preset time period; according to the online time record of the game player after logging in each game, calculating and obtaining the sum of the online time of each game in a preset time period; the characteristics of the active dimension of the gamer then include the name of the game played by the gamer, the total number of days the game was played during the preset time period, the sum of the online time durations of the games during the preset time period, and so forth.
Step S240, extracting features of social dimensions of the game player according to social data of the game player in the database.
Wherein the social data of the game player comprises: game player's game friends in each type of game, affinity values with each game friend, interaction data with the game friend, and the like, the interaction data including sharing, praying, gifting, and the like. Extracting features of social dimensions of the game player according to social data of the game player in the database, such as: obtaining game friends of a game player in each type of game and affinity values of the game friends from a database; determining the number of friends of a game player according to the game friends of the game player in each type of game; determining the total intimacy value of the game player and each game friend according to the game friends of the game player in each game and the intimacy value of the game player and each game friend; the number of friends of the game player and the total affinity value of the game player and each game friend are features of the social dimension of the game player. The social dimension characteristics of the game player can also comprise friend interaction information of the game player, interaction data of the game player and the game friends are obtained from a database, and the interaction data comprise sharing, praying and giving; and extracting characteristics according to the interaction data of the game player and the game friends to obtain friend interaction information of the game player.
Determining the total intimacy value of the game player and each game friend according to the game friends of the game player in each game and the intimacy value of the game player and each game friend, wherein the total intimacy value of the game player and each game friend is as follows: when a game player is in a friend relationship with a game friend involving multiple suits in a single game, such as: players a and B may be friends in several of the zone gowns and then sum up the affinity values of players a and B in several of the zone gowns. When a game player and a certain game friend relate to a plurality of games, unique identification ids of the game player are unified from a single game to one identification id related to QQ or WeChat. Then, the affinity values of different games need to be mapped between 0 and 1, and extremely abnormal values are removed, so that the affinity median value of each game is calibrated to the similar value, and the affinity values of the game players and the game friends are summed again based on QQ or identification id related to a WeChat platform.
In one embodiment, the key user identification method further comprises: processing the features of the social dimension of the game player by adopting a graph embedding method and a graph representation learning method to obtain network embedded features of the game player; network embedded features of gamers are added as features of gamers' social dimensions.
The graph embedding method refers to a typical method in graph representation learning, and features on the graph are represented in the form of compressed vectors or matrixes. The graph representation learning method is a method of representing features on a graph as other forms. After the characteristics of the social dimension of the game player are processed by adopting the graph embedding method and the graph representation learning method, the obtained network embedded characteristics of the game player are also used as a part of the characteristics of the social dimension of the game player, so that the characteristics of the social dimension of the game player are further supplemented, the user portraits of the game player are enriched, and the accuracy of identifying key users corresponding to the game is improved.
And step S260, extracting the characteristics of each game according to the game knowledge graph to obtain the game characteristics of each game.
Wherein, the game knowledge graph is a structure for storing and representing data, and the core is a triplet relation, for example: QQ galloping- & gtgames belonging to the class of racing; QQ galloping → developer → Tencer. The ternary form is called a map form, is convenient to store and can automatically obtain complex knowledge.
In one embodiment, the step of extracting features of each game according to the game knowledge graph to obtain game features of each game includes any one or more of the following steps:
First kind: constructing n-dimensional vectors of each game according to the dimensions of the subsystems, and acquiring game features of each game; second kind: according to the game knowledge graph of each game, representing each game into a vector embedded based on a network, and obtaining the game characteristics of each game; third kind: and finding and arranging similar games of the current game through a related similarity calculation method, and obtaining game characteristics of each game.
Wherein the subsystem is determined by dividing according to different games. The games are represented as network-embedded vectors, which are represented as network-embedded vectors by a network embedding method, which may be word2vec (a related model used to generate word vectors), deep walk (a new method of learning the representation of nodes in a network), node2vec (a model used to generate node vectors in a network), gcn (a graph roll-up network), and so forth. The similarity calculation method may be cosine similarity calculation, euclidean distance size, or the like. The method is characterized in that the similar games of the current game are found and arranged through a related similarity calculation method, the game characteristics of each game are obtained, and on the basis of finding and arranging the similar games of the current game through the related similarity calculation method, the similar games can be further diffused to the similar games of the similar games based on the similarity calculation method, so that the range of the similar games is expanded, and the game characteristics of each game are obtained.
The feature extraction is performed on each game according to the game knowledge graph to obtain the game feature of each game, which can be obtained by adopting any one feature extraction mode of the first type, the second type and the third type, or by arbitrarily selecting two types of the first type, the second type and the third type to perform feature extraction respectively to obtain the game feature of each game, or by adopting the first type, the second type and the third type to perform feature extraction respectively to obtain the game feature of each game.
And step S280, performing feature cross combination according to the features of the active dimension, the features of the social dimension and the game features of each game to obtain the combined features of the game players after multi-dimension cross.
Wherein, the feature of the active dimension of the game player, the feature of the social dimension and the game feature of each game are combined in a feature cross way, such as: according to the combination of the characteristics of the active dimension of the game player and the characteristics of the social dimension, the characteristics of the game player in the active and social dimension can be obtained; the feature cross combination is carried out according to the feature of the active dimension of the game player and the game feature of each game, so that the game player can show which games are active, and the game feature of the game is what, namely the feature of the dimension of the active+game feature; the social dimension characteristics of the game player and the game characteristics of each game can show the social capability of the game player, what the game characteristics of the game played by the game player are, and the characteristics of the game player in the dimension of the social plus game characteristics can be obtained; the feature of the active dimension of the game player, the feature of the social dimension and the game feature of each game are combined, the similarity between the active games of the game player can be determined according to the feature of the active dimension of the player and the game feature of each game, the social capability of the game player in the similar games can be further reflected, and the feature of the game player in the social dimension of the similar games can be obtained. The multi-modal characteristics which are richer and more perfect for the target crowd are obtained through further combination and extraction.
Step S300, carrying out cluster analysis according to the active dimension characteristics, the social dimension characteristics and the multi-dimension crossed combined characteristics of the game player to obtain the user image of the game player in the preset dimension.
The clustering analysis can be to adopt a K-MEANS algorithm, a K-MEDOIDS algorithm, a CLARANS algorithm and the like, and perform clustering analysis from different dimensions according to the active dimension characteristics, the social dimension characteristics and the multi-dimension crossed combined characteristics of the game player, so that the game player is classified in different dimensions, and user portraits are performed for the game player in different dimensions according to classification results of the game player in different dimensions. The preset dimension can be determined according to the characteristics of the game player and the service requirement, and user portraits are carried out on the game player in different dimensions based on the characteristics of the game player.
In one embodiment, the step of performing cluster analysis according to the active dimension feature, the social dimension feature, and the multi-dimension intersected combined feature of the game player to obtain a user portrait of the game player in a preset dimension includes:
respectively carrying out cluster analysis according to the active dimension characteristics, the social dimension characteristics and the multi-dimension crossed combined characteristics of the game player to obtain user figures of the game player in each dimension; according to the characteristics of the active dimension of the game player, obtaining the game characteristics corresponding to the game logged in by the game player; clustering is carried out according to game characteristics corresponding to games passed by the game player, and game preference portraits of the game player are obtained.
The method comprises the steps of respectively carrying out cluster analysis according to the characteristics of the active dimension, the characteristics of the social dimension and the combined characteristics after multi-dimensional intersection of the game player, wherein the cluster analysis is carried out on the dimensions formed by the active dimension, the social dimension and the combined characteristics after the intersection of the game player. The user portraits of the game player in each dimension comprise the user portraits of the game player in the active dimension, the user portraits of the social dimension and the user portraits of the dimension formed by the combined characteristics after each intersection, namely: user portraits of the liveness + social dimension, user portraits of the liveness + game dimension, user portraits of the dimension of the social + game feature, and user portraits of the social dimension of a similar game. According to the characteristics of the active dimension of the game player, the game characteristics of the game are further acquired according to the game logged in by the game player, and clustering is carried out according to the game characteristics of the game logged in by the game player, so that the characteristics of the game played by the game player can be determined, and the game preference portrait of the game player is obtained.
The game players can be classified according to the characteristics of the active dimension of the game players, different categories respectively represent the liveness of different degrees, user portraits are carried out for the game players of each category, and the user portraits of the active dimension of the game players of the same category are the same. The game players can be classified according to the dimension combination characteristics formed by the combination characteristics after the intersection of the game players, different categories represent the degrees of the game players, user figures are carried out for the game players of the different categories, and the user figures of the game players of the same category are the same.
The game players can be classified according to the characteristics of the social dimension of the game players, different categories respectively represent social degrees of different degrees, user portraits are carried out for the game players of the different categories, and the user portraits of the social dimension of the game players of the same category are the same. Further, the relationship chains of the players can be clustered based on category games (such as categories of MOBA, RPG, racing games, leisure games and the like), similar games (each category of games forms a feature vector based on a game knowledge graph, cosine similarity (or other similarity) of the feature vectors of different games is calculated, the similar games are determined according to the cosine similarity from big to small), the relationship chains of the players are clustered under different scales of a game big disc and the like, multiple relationship chains formed by multiple games are obtained, social relationships of the game players under multiple ranges of sizes are obtained, and user figures of social dimensions of the game players under different ranges are constructed.
In one scene, taking the example that a QQ flyer hand game needs to be pulled by a new user, the class of the QQ flyer hand game is a racing game, a user portrait of social dimension can be constructed for a game player based on the racing game, and the social degree of the game player in the racing game can be reflected; similarly, the similar games based on the QQ flyer hand game construct user portraits of social dimension for game players, so that the social degree of the game players in the similar games of the QQ flyer hand game can be reflected.
Step S320, according to the received attention dimension of the identification condition, the corresponding key user is identified from the game player based on the user portrait of the game player in the preset dimension.
The identification condition comprises one or more of information of the number of key users, attention dimension, user source (QQ number/micro signal/mobile phone number), identification of a target game, seed users and the like to be acquired. The target game is a game requiring acquisition of a key user, and the identification of the target game may be a game name. The dimension of interest refers to which one or more of the preset dimensions the game player is focused on when identifying the key user. The key users refer to the users with the core inside and outside the game, the greatest influence and the strongest calling force, and by introducing the users into the game, the users can share the information added into the game to other people so as to drive other people to also add into the game. The game advertisement can be put into the key users through the identified key users, and the game advertisement is used for user pull-up, user recall and the like of the game service, and is also suitable for marketing operation activities for encouraging the users to share and spread in social ways.
In one embodiment, referring to fig. 3, according to the received attention dimension of the recognition condition, the step of recognizing the corresponding key user from the game player based on the user representation of the game player in the preset dimension includes:
Step S322, according to the received identification conditions, the attention dimension of the key user is determined.
Step S324, a user portrait of the game player in the concerned dimension is obtained.
Step S326, the game player is filtered according to the received identification conditions and the user images of the game player in the attention dimension to obtain key users.
Wherein the dimension of interest may be one or more of the preset dimensions. Filtering the game player according to the received identification conditions and the user images of the game player in the attention dimension to obtain key users, such as: assuming that the key users of the target game A are acquired, the key users need to have liveness of 90 in the liveness dimension, the attention dimension of the key users is identified as the liveness dimension, the user portraits of the game players in the attention dimension are acquired, and the game players with liveness of 90 are selected from the game players as the key users. And when the identification conditions are other than the attention dimension, further screening and obtaining the key users according to the other conditions.
In one embodiment, the step of filtering the game player to obtain key users based on the received identification conditions and the user images of the game player in the dimension of interest comprises:
Filtering the game player according to the received identification conditions and the user images of the game player in the attention dimension, and filtering the game player to obtain a preselected user; and analyzing the probability of playing the target game by the preselected user according to the user portrait of the preselected user in the preset dimension, and determining the user with the probability of playing the target game larger than the preset value in the preselected user as the key user.
The method comprises the steps of analyzing the probability of playing the target game by the preselected user in the user portrait of the preset dimension, namely inputting the user portrait of the preselected user in the preset dimension into a key user identification model, analyzing the probability of playing the target game by the preselected user through the key user identification model, dividing the key user identification model into a training set and a prediction set according to the time sequence according to the user data of collected game players, and predicting the game players on the target game based on whether the game players log in the similar game or not, wherein the key user identification model can be predicted by adopting a tree model classification algorithm (random forest, xgboost, multi-layer perceptron, logistic Regression and the like), and a Deep neural network algorithm (Deep Interest Network, wide Deep, network Embedding, GCN and the like). The probability of playing the target game by the preselected user is analyzed by the user portrait of the preselected user in the preset dimension, and the user with the probability of playing the target game larger than the preset value in the preselected user is determined as the key user, so that the accuracy of the identification of the key user can be further improved.
In one embodiment, the step of identifying the corresponding key user from the game player based on the user representation of the game player in the preset dimension according to the received dimension of interest of the identification condition comprises:
when the received identification condition provides the seed user, acquiring a user image of the seed user in a preset dimension; according to the user portrait of the seed user in the preset dimension, determining the attention dimension and the user portrait corresponding to the attention dimension; a user having a similar user profile to the seed user is determined from the game player as a key user based on the attention dimension and the user profile corresponding to the attention dimension.
Where a seed user refers to a user who has played a target game. According to the user portraits of the seed users in the preset dimension, the user portraits of the seed users in one or more dimensions can be obtained, so that the dimension of the same user portraits is determined to be the concerned dimension, the user portraits corresponding to the concerned dimension are determined, and according to the concerned dimension and the user portraits corresponding to the concerned dimension, the users with similar user portraits in the concerned dimension and the seed users can be screened out from all game players to serve as key users.
When the seed users are provided in the identification conditions, users which are in friend relation with the seed users can be screened out from all game players according to social characteristics of the game players, and whether the users can be key users is determined according to the fact that the total intimacy value of the game players and the seed users reaches a preset value or the number of key users which are the seed users in the game friends of the game players. The user is easily influenced by friends to play a game, so that the key user is determined according to the seed user, and the accuracy of key user identification can be improved.
According to the key user identification method, the characteristics of the active dimension and the characteristics of the social dimension of the game player are extracted according to the active data and the social data of the game player in the database; obtaining game characteristics of each game according to the game knowledge graph; performing feature cross combination on the features of the active dimension of the game player, the features of the social dimension and the game features of each game to obtain combined features of the game player after multi-dimensional cross; performing cluster analysis according to the combination characteristics of the game players after multi-dimensional intersection to obtain user images of the game players in preset dimensions; according to different games, corresponding identification conditions are determined, and key users corresponding to the games are identified from the game players according to the attention dimension of the identification conditions and the user portraits of the game players in preset dimensions, so that the accuracy of identifying the key users for different games is improved.
In one embodiment, a key user identification method is further described for a hundred million level game player based on tdw (Tencent Distributed Warehouse: tencent distributed data warehouse) offline data, implemented by SQL (structured query language (Structured Query Language)) run or spark (fast general purpose computing engine designed for large scale data processing) run on:
the method comprises the steps of integrating and collecting recently (30 days, 7 days or 14 days and the like) logged-in game information, logged-in day information and logged-in time length information of a game player and the like from a plurality of data sources (databases), obtaining the characteristics of the active dimension of the game player, clustering according to the characteristics of the active dimension of the game player, and obtaining the user portrait of the game player in the active dimension. According to the feature of the active dimension of the game player, the game features of the game are further acquired according to the game logged in by the game player, and clustering is carried out according to the game features corresponding to the game logged in by the game player, so that the characteristics of the game played by the game player can be determined, and the game preference portrait of the game player is acquired.
Integrating and collecting game friends of a game user in each type of game from a plurality of data sources (databases), and determining the number of friends of a game player according to the affinity value of the game user and each game friend in each type of game; according to game friends of a game player in each game and the affinity value of the game player and each game friend; thereby obtaining characteristics of the social dimension of the game player; further, the characteristics of the social dimension of the game player are processed by adopting a graph embedding method and a graph representation learning method, so as to obtain network embedded characteristics of the game player; network embedded features of gamers are added as features of gamers' social dimensions. Clustering is carried out according to the characteristics of the social dimension of the game player, and the user portraits of the game player in the social dimension are obtained.
Further, the relationship chains of the players can be clustered based on category games (such as categories of MOBA, RPG, racing games, leisure games and the like), similar games (each category of games forms a feature vector based on a game knowledge graph, cosine similarity (or other similarity) of the feature vectors of different games is calculated, the similar games are determined according to the cosine similarity from big to small), the relationship chains of the players are clustered under different scales of a game big disc and the like, multiple relationship chains formed by multiple games are obtained, social relationships of the game players under multiple ranges of sizes are obtained, and user figures of social dimensions of the game players under different ranges are constructed. Further, each game can be sequentially used as a target game, the users who have played the target game are used as seed users, diffusion is carried out on the basis of the seed set users under different scales of the class game, the similar game, the game big disc and the like, and during diffusion, secondary friends of the seed players are obtained based on affinity weight sequencing between the two players, so that user figures of the game players based on social dimensions of the seed users are constructed. If the target game does not provide a seed set user, then no consideration is given.
Constructing n-dimensional vectors of each game according to the dimensions of the subsystems, and acquiring game features of each game; according to the game knowledge graph of each game, representing each game into a vector embedded based on a network, and obtaining the game characteristics of each game; and finding and arranging similar games of the current game through a related similarity calculation method, and obtaining game characteristics of each game. The similarity calculation method can be further based on that the similar games are further diffused to the similar games of the similar games, the range of the similar games is expanded, and the game characteristics of each game are obtained.
Referring to fig. 4, features of the game player in the active+social dimension may be obtained by combining features of the game player in the active dimension with features of the social dimension; the feature cross combination is carried out according to the feature of the active dimension of the game player and the game feature of each game, so that the game player can show which games are active, and the game feature of the game is what, namely the feature of the dimension of the active+game feature; the social dimension characteristics of the game player and the game characteristics of each game can show the social capability of the game player, what the game characteristics of the game played by the game player are, and the characteristics of the game player in the dimension of the social plus game characteristics can be obtained; combining the active dimension characteristics of the game player, the social dimension characteristics and the game characteristics of each game, and determining the similarity between the active games of the game player according to the active dimension characteristics of the player and the game characteristics of each game; further, the social ability of the game player in the similar game can be reflected, and the characteristics of the game player in the social dimension of the similar game can be obtained. Clustering according to the characteristics of the game player in the dimension of liveness and social contact to obtain user portraits of the dimension of liveness and social contact; clustering according to the feature of the dimension of the active+game feature to obtain the user portrait of the active+game dimension; clustering according to the feature of the game player in the dimension of the social contact + game feature to obtain a user portrait of the game player in the dimension of the social contact + game feature; clustering is carried out according to the characteristics of the game players in the social dimension of the similar game, and user portraits of the social dimension of the similar game are obtained.
When a key user is required to be acquired in the scenes of a user pull-up, a user recall, a marketing operation activity encouraging the user to socially share and spread, and the like of the game service, an identification condition is determined according to the scenes of the game service, and a key user identification device identifies the corresponding key user from game players based on the user portraits of the game players in the preset dimension according to the received attention dimension of the identification condition.
The key user identifying device determines the step of identifying the key user according to the received identifying condition, and the identifying condition comprises the following steps: when the attention dimension and the portrait condition corresponding to the attention dimension take the attention dimension as an example, the liveness degree of the liveness dimension reaches 90, and the liveness degree reaches 90 as the portrait condition), the user portrait of the game player in the attention dimension is acquired according to the attention dimension of the identified key user, and the key user is obtained by filtering the game player according to the portrait condition corresponding to the attention dimension and the user portrait of the game player in the attention dimension in the received identification condition. If the key users need to be further filtered, the game player can be filtered according to the received identification conditions and the user images of the game player in the attention dimension, and the key users are obtained from the filtering of the game player and serve as preselected users; inputting the user portrait of the preselected user in the preset dimension into a key user identification model to analyze the probability of playing the target game of the preselected user, and determining the user with the probability of playing the target game larger than the preset value in the preselected user as the key user.
When the identification condition includes: when the target game is identified, the probability of playing the target game by the game player can be analyzed by inputting the user portrait of the game player in the preset dimension into the key user identification model, and the user with the probability of playing the target game by the game player being greater than the preset value is determined to be the key user.
According to the key user identification method, based on the fact that user portraits are carried out on game players in multiple dimensions, portraits for judging whether the game players are key users or not can be provided for different games, identification conditions can be determined for new games, so that the key users can be acquired for user updating, identification conditions can be determined for old games, and therefore the key users can be acquired for user updating; the identification conditions can be determined for popular games in the civilian direction, so that key users are obtained for updating, and the identification conditions can be determined for popular games in the public, so that key users are obtained for updating; the identification conditions may be determined for a certain game category to obtain key users for user pull-up, or for similar games to obtain key users for user pull-up, etc.
Fig. 2-3 are flow diagrams of a method of key user identification in one embodiment. It should be understood that, although the steps in the flowcharts of fig. 2-3 are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited to the order of execution unless explicitly recited herein, and the steps may be executed in other orders. Moreover, at least some of the steps in fig. 2-3 may include multiple sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, nor do the order in which the sub-steps or stages are performed necessarily occur sequentially, but may be performed alternately or alternately with at least a portion of the sub-steps or stages of other steps or steps.
In one embodiment, referring to fig. 5, a key subscriber identification device includes: active feature extraction module 310, social feature extraction module 320, game feature extraction module 330, feature cross-combination module 340, clustering module 350, and key user identification module 360.
An active feature extraction module 310, configured to extract features of active dimensions of a game player according to active data of the game player in the database;
the social feature extraction module 320 is configured to extract features of social dimensions of the game player according to social data of the game player in the database;
the game feature extraction module 330 is configured to perform feature extraction on each game according to the game knowledge graph, so as to obtain game features of each game;
the feature cross combination module 340 is configured to perform feature cross combination according to the feature of the active dimension of the game player, the feature of the social dimension, and the game feature of each game, so as to obtain a combined feature of the game player after multi-dimensional cross;
the clustering module 350 is configured to perform cluster analysis according to the active dimension features, the social dimension features, and the multi-dimension crossed combined features of the game player, so as to obtain a user image of the game player in a preset dimension;
The key user identification module 360 is configured to identify, from among game players, corresponding key users based on user portraits of the game players in preset dimensions according to the received attention dimensions of the identification conditions.
In one embodiment, the key subscriber identity module 360 is further configured to: according to the received identification conditions, determining the attention dimension of the identified key user; acquiring a user portrait of a game player in a concerned dimension; and filtering the game player according to the received identification conditions and the user images of the game player in the attention dimension to obtain key users.
In one embodiment, the key subscriber identity module 360 is further configured to: filtering the game player according to the received identification conditions and the user images of the game player in the attention dimension, and filtering the game player to obtain a preselected user; and analyzing the probability of playing the target game by the preselected user according to the user portrait of the preselected user in the preset dimension, and determining the user with the probability of playing the target game larger than the preset value in the preselected user as the key user.
In one embodiment, the key subscriber identity module 360 is further configured to: when the received identification condition provides the seed user, acquiring a user image of the seed user in a preset dimension; according to the user portrait of the seed user in the preset dimension, determining the attention dimension and the user portrait corresponding to the attention dimension; a user having a similar user profile to the seed user is determined from the game player as a key user based on the attention dimension and the user profile corresponding to the attention dimension.
In one embodiment, the clustering module 350 is further configured to: respectively carrying out cluster analysis according to the active dimension characteristics, the social dimension characteristics and the multi-dimension crossed combined characteristics of the game player to obtain user figures of the game player in each dimension; according to the characteristics of the active dimension of the game player, obtaining the game characteristics of the game logged in by the game player; clustering is carried out according to game characteristics of games played by the game players, and game preference portraits of the game players are obtained.
In one embodiment, as shown in fig. 6, the key user identification device further includes a social feature processing module 370, configured to process the features of the social dimension of the game player by using a graph embedding method and a graph representation learning method, so as to obtain network embedded features of the game player; network embedded features of gamers are added as features of gamers' social dimensions.
In one embodiment, the game feature extraction module 330 is further configured to: constructing n-dimensional vectors of each game according to the dimensions of the subsystems, and acquiring game features of each game; according to the game knowledge graph of each game, representing each game into a vector embedded based on a network, and obtaining the game characteristics of each game; and finding and arranging similar games of the current game through a related similarity calculation method, and obtaining game characteristics of each game.
According to the key user identification device, the active feature extraction module 310 and the social feature extraction module 320 are used for extracting the features of the active dimension and the features of the social dimension of the game player according to the active data and the social data of the game player in the database respectively; the game feature extraction module 330 obtains game features of each game according to the game knowledge graph; the feature cross combination module 340 performs feature cross combination on the features of the active dimension, the features of the social dimension and the game features of each game to obtain combined features of the game player after multi-dimensional cross; clustering module 350 performs cluster analysis according to the combined characteristics of the game players after multi-dimensional intersection to obtain user images of the game players in preset dimensions; therefore, the key user identification module 360 determines corresponding identification conditions according to different games, identifies key users corresponding to the games from the game players according to the attention dimension of the identification conditions and the user portraits of the game players in preset dimensions, and improves the accuracy of identifying the key users for different games.
FIG. 7 illustrates an internal block diagram of a computer device in one embodiment. The computer device may be specifically the terminal 110 or the server 120 in fig. 1. As shown in fig. 7, the computer device includes a processor, a memory, a network interface, an input device, and a display screen connected by a system bus. The memory includes a nonvolatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system and may also store a computer program which, when executed by a processor, causes the processor to implement a key user identification method. The internal memory may also have stored therein a computer program which, when executed by the processor, causes the processor to perform the method of critical user identification. The display screen of the computer equipment can be a liquid crystal display screen or an electronic ink display screen, the input device of the computer equipment can be a touch layer covered on the display screen, can also be keys, a track ball or a touch pad arranged on the shell of the computer equipment, and can also be an external keyboard, a touch pad or a mouse and the like.
It will be appreciated by those skilled in the art that the structure shown in FIG. 7 is merely a block diagram of some of the structures associated with the present inventive arrangements and is not limiting of the computer device to which the present inventive arrangements may be applied, and that a particular computer device may include more or fewer components than shown, or may combine some of the components, or have a different arrangement of components.
In one embodiment, the key user identification means provided by the present application may be implemented in the form of a computer program which is executable on a computer device as shown in fig. 7. The memory of the computer device may store various program modules that make up the key user identification means, such as the active feature extraction module 310, the social feature extraction module 320, the game feature extraction module 330, the feature cross-combination module 340, the clustering module 350, and the key user identification module 360 shown in fig. 5. The computer program of each program module causes the processor to carry out the steps of the method for identifying a key user according to each embodiment of the present application described in the present specification.
For example, the computer device shown in FIG. 7 may extract features of the active dimension of a game player by executing the active data of the game player in the database through the active feature extraction module 310 in the key user identification device shown in FIG. 5. The computer device may extract features of the social dimension of the game player from social data of the game player in the database by the social feature extraction module 320. The computer device may perform feature extraction on each game according to the game knowledge graph through the game feature extraction module 330 to obtain game features of each game. The computer device may perform feature cross-combining according to the active dimension features, the social dimension features, and the game features of each game of the game player through feature cross-combining module 340 to obtain a multi-dimensionally crossed combined feature of the game player. The computer device may perform a cluster analysis according to the active dimension features, the social dimension features, and the multi-dimensionally intersected combined features of the game player through the clustering module 350 to obtain a user image of the game player in a preset dimension. The computer device may execute the dimension of interest according to the received recognition conditions via the key user recognition module 360 to identify a corresponding key user from the game player based on the user representation of the game player in the preset dimension.
In one embodiment, a computer device is provided comprising a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the above-described critical user identification method. The steps of the key user identification method herein may be the steps of the key user identification method of the above embodiments.
In one embodiment, a computer readable storage medium is provided, storing a computer program which, when executed by a processor, causes the processor to perform the steps of the above-described key user identification method. The steps of the key user identification method herein may be the steps of the key user identification method of the above embodiments.
Those skilled in the art will appreciate that all or part of the processes in the methods of the above embodiments may be implemented by a computer program for instructing relevant hardware, where the program may be stored in a non-volatile computer readable storage medium, and where the program, when executed, may include processes in the embodiments of the methods described above. Any reference to memory, storage, database, or other medium used in embodiments provided herein may include non-volatile and/or volatile memory. The nonvolatile memory can include Read Only Memory (ROM), programmable ROM (PROM), electrically Programmable ROM (EPROM), electrically Erasable Programmable ROM (EEPROM), or flash memory. Volatile memory can include Random Access Memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms such as Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double Data Rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), memory bus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), among others.
The technical features of the above embodiments may be arbitrarily combined, and all possible combinations of the technical features in the above embodiments are not described for brevity of description, however, as long as there is no contradiction between the combinations of the technical features, they should be considered as the scope of the description.
The foregoing examples illustrate only a few embodiments of the application and are described in detail herein without thereby limiting the scope of the application. It should be noted that it will be apparent to those skilled in the art that several variations and modifications can be made without departing from the spirit of the application, which are all within the scope of the application. Accordingly, the scope of protection of the present application is to be determined by the appended claims.

Claims (16)

1. A method of key user identification, comprising:
extracting the characteristics of the active dimension of the game player according to the active data of the game player in the database;
extracting the characteristics of the social dimension of the game player according to the social data of the game player in the database;
extracting features of each game according to the game knowledge graph to obtain game features of each game; the game knowledge graph is a data structure used for representing the relation between each game, game class and developer, and the game features are features of game dimension;
Performing feature cross combination according to the features of the active dimension of the game player, the features of the social dimension and the game features of each game to obtain combined features of the game player after multi-dimensional cross;
performing cluster analysis according to the active dimension characteristics, the social dimension characteristics and the multi-dimension crossed combined characteristics of the game player to obtain a user image of the game player in a preset dimension; the preset dimension is the active dimension, the social dimension and the game dimension respectively, and a combination dimension among the active dimension, the social dimension and the game dimension;
when a seed user is provided in the received identification condition, acquiring a user portrait of the seed user in a preset dimension; according to the user portrait of the seed user in the preset dimension, determining the concerned dimension and the user portrait corresponding to the concerned dimension; determining users with similar user portraits to the seed users from the game players according to the attention dimension and the user portraits corresponding to the attention dimension as key users; the dimension of interest is at least one dimension of the preset dimensions that is focused on the game player when the key user is identified.
2. The method according to claim 1, wherein the method further comprises:
according to the received identification conditions, determining the attention dimension of the identified key user;
acquiring a user portrait of the game player in a concerned dimension;
and filtering the game player according to the received identification conditions and the user image of the game player in the attention dimension to obtain a key user.
3. The method of claim 2, wherein the step of filtering the game player for key users based on the received identification conditions and the user image of the game player in a dimension of interest comprises:
filtering the game player according to the received identification conditions and the user images of the game player in the concerned dimension, and filtering the game player to obtain a preselected user;
and analyzing the probability of playing the target game by the preselected user according to the user portrait of the preselected user in the preset dimension, and determining the user with the probability of playing the target game larger than a preset value in the preselected user as a key user.
4. The method according to claim 1, wherein the method further comprises:
When a seed user is provided in the received identification condition, screening users which are in friend relation with the seed user from all game players according to the social characteristics of the game players;
and determining a key user according to the fact that the total intimacy value of the game player and the seed user reaches a preset value.
5. The method according to claim 1, wherein the step of performing cluster analysis according to the active dimension feature, the social dimension feature, and the multi-dimensionally intersected combined feature of the game player to obtain the user representation of the game player in the preset dimension comprises:
respectively carrying out cluster analysis according to the active dimension characteristics, the social dimension characteristics and the multi-dimension crossed combined characteristics of the game player to obtain user figures of the game player in each dimension;
according to the characteristics of the active dimension of the game player, obtaining the game characteristics of the game logged in by the game player;
and clustering according to game features of the games played by the game players to obtain game preference portraits of the game players.
6. The method of claim 1, wherein the method further comprises:
Processing the characteristics of the social dimension of the game player by adopting a graph embedding method and a graph representation learning method to obtain network embedded characteristics of the game player;
network embedded features of the gamer are added as features of the gamer's social dimension.
7. The method of claim 1, wherein the step of extracting features of each game according to the game knowledge graph to obtain game features of each game comprises any one or more of the following steps:
first kind: constructing n-dimensional vectors of each game according to the dimensions of the subsystems, and acquiring game features of each game;
second kind: according to the game knowledge graph of each game, representing each game into a vector embedded based on a network, and obtaining the game characteristics of each game;
third kind: and finding and arranging similar games of the current game through a related similarity calculation method, and obtaining game characteristics of each game.
8. A key subscriber identification device, comprising:
the active feature extraction module is used for extracting the features of the active dimension of the game player according to the active data of the game player in the database;
the social feature extraction module is used for extracting the features of the social dimension of the game player according to the social data of the game player in the database;
The game feature extraction module is used for extracting features of each game according to the game knowledge graph to obtain game features of each game; the game feature is a feature of a game dimension;
the feature cross combination module is used for carrying out feature cross combination according to the features of the active dimension, the features of the social dimension and the game features of each game of the game player to obtain combined features of the game player after multi-dimension cross;
the clustering module is used for carrying out clustering analysis according to the active dimension characteristics, the social dimension characteristics and the multi-dimension crossed combined characteristics of the game player to obtain user images of the game player in the preset dimension; the preset dimension is the active dimension, the social dimension and the game dimension respectively, and a combination dimension among the active dimension, the social dimension and the game dimension;
the key user identification module is used for acquiring a user portrait of the seed user in a preset dimension when the seed user is provided in the received identification condition; according to the user portrait of the seed user in the preset dimension, determining the concerned dimension and the user portrait corresponding to the concerned dimension; determining users with similar user portraits to the seed users from the game players according to the attention dimension and the user portraits corresponding to the attention dimension as key users; the dimension of interest is at least one dimension of the preset dimensions that is focused on the game player when the key user is identified.
9. The apparatus of claim 8, wherein the key user identification module is further configured to determine a dimension of interest identifying a key user based on the received identification condition; acquiring a user portrait of the game player in a concerned dimension; and filtering the game player according to the received identification conditions and the user image of the game player in the attention dimension to obtain a key user.
10. The apparatus of claim 9, wherein the key user identification module is further configured to filter the game player based on the received identification condition and a user image of the game player in the dimension of interest, filtering a preselected user from the game player; and analyzing the probability of playing the target game by the preselected user according to the user portrait of the preselected user in the preset dimension, and determining the user with the probability of playing the target game larger than a preset value in the preselected user as a key user.
11. The apparatus of claim 8, wherein the key user identification module is further configured to screen all game players for users that are in a friend relationship with the seed user based on social characteristics of the game players when the seed user is provided in the received identification condition; and determining a key user according to the fact that the total intimacy value of the game player and the seed user reaches a preset value.
12. The apparatus of claim 8, wherein the clustering module is further configured to perform cluster analysis according to the active dimension feature, the social dimension feature, and the multi-dimensionally intersected combination feature of the game player, to obtain a user representation of the game player in each dimension; according to the characteristics of the active dimension of the game player, obtaining the game characteristics of the game logged in by the game player; and clustering according to game features of the games played by the game players to obtain game preference portraits of the game players.
13. The apparatus of claim 8, wherein the apparatus further comprises:
the social feature processing module is used for processing the features of the social dimension of the game player by adopting the graph embedding device and the graph representation learning device to obtain network embedding features of the game player; network embedded features of the gamer are added as features of the gamer's social dimension.
14. The apparatus of claim 8, wherein the game feature extraction module is further configured to construct an n-dimensional vector for each game according to a plurality of subsystem dimensions, and obtain a game feature for each game; according to the game knowledge graph of each game, representing each game into a vector embedded based on a network, and obtaining the game characteristics of each game; and finding and arranging similar games of the current game through a related similarity computing device, and obtaining game characteristics of each game.
15. A computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of the method of any one of claims 1 to 7.
16. A computer device comprising a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the steps of the method of any of claims 1 to 7.
CN202010092848.7A 2020-02-14 2020-02-14 Key user identification method, key user identification device, readable storage medium and computer equipment Active CN113268589B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010092848.7A CN113268589B (en) 2020-02-14 2020-02-14 Key user identification method, key user identification device, readable storage medium and computer equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010092848.7A CN113268589B (en) 2020-02-14 2020-02-14 Key user identification method, key user identification device, readable storage medium and computer equipment

Publications (2)

Publication Number Publication Date
CN113268589A CN113268589A (en) 2021-08-17
CN113268589B true CN113268589B (en) 2023-09-22

Family

ID=77227353

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010092848.7A Active CN113268589B (en) 2020-02-14 2020-02-14 Key user identification method, key user identification device, readable storage medium and computer equipment

Country Status (1)

Country Link
CN (1) CN113268589B (en)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113648659B (en) * 2021-08-20 2023-09-26 腾讯科技（深圳）有限公司 Method and related device for determining user liveness
CN115168740B (en) * 2022-09-06 2022-11-15 网娱互动科技(北京)股份有限公司 Method and system for generating marketing task based on big data analysis

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105477860A (en) * 2015-12-22 2016-04-13 北京奇虎科技有限公司 Game activity recommending method and device
CN105574110A (en) * 2015-12-14 2016-05-11 北京奇虎科技有限公司 Intelligent game recommending method and device
CN107292465A (en) * 2016-03-31 2017-10-24 阿里巴巴集团控股有限公司 User's evaluation method, device and equipment
KR20190097874A (en) * 2018-02-13 2019-08-21 이승욱 Apparatus for game recommendation service and method thereof
CN110489453A (en) * 2019-07-02 2019-11-22 广东工业大学 User's game real-time recommendation method and system based on big data log analysis
CN110674394A (en) * 2019-08-20 2020-01-10 腾讯科技（深圳）有限公司 Knowledge graph-based information recommendation method and device and storage medium

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105574110A (en) * 2015-12-14 2016-05-11 北京奇虎科技有限公司 Intelligent game recommending method and device
CN105477860A (en) * 2015-12-22 2016-04-13 北京奇虎科技有限公司 Game activity recommending method and device
CN107292465A (en) * 2016-03-31 2017-10-24 阿里巴巴集团控股有限公司 User's evaluation method, device and equipment
KR20190097874A (en) * 2018-02-13 2019-08-21 이승욱 Apparatus for game recommendation service and method thereof
CN110489453A (en) * 2019-07-02 2019-11-22 广东工业大学 User's game real-time recommendation method and system based on big data log analysis
CN110674394A (en) * 2019-08-20 2020-01-10 腾讯科技（深圳）有限公司 Knowledge graph-based information recommendation method and device and storage medium

Also Published As

Publication number Publication date
CN113268589A (en) 2021-08-17

Similar Documents

Publication Publication Date Title
CN110012356B (en) Video recommendation method, device and equipment and computer storage medium
CN108427708B (en) Data processing method, data processing apparatus, storage medium, and electronic apparatus
CN111177473B (en) Personnel relationship analysis method, device and readable storage medium
CN111773732B (en) Target game user detection method, device and equipment
CN111898031A (en) Method and device for obtaining user portrait
CN113971527A (en) Data risk assessment method and device based on machine learning
CN113268589B (en) Key user identification method, key user identification device, readable storage medium and computer equipment
CN114223012A (en) Push object determination method and device, terminal equipment and storage medium
Concolato et al. Data science: A new paradigm in the age of big-data science and analytics
CN110737811A (en) Application classification method and device and related equipment
CN112989179B (en) Model training and multimedia content recommendation method and device
US11481580B2 (en) Accessible machine learning
CN116764236A (en) Game prop recommending method, game prop recommending device, computer equipment and storage medium
CN114519508A (en) Credit risk assessment method based on time sequence deep learning and legal document information
CN110209704B (en) User matching method and device
CN116523293A (en) User risk assessment method based on fusion behavior flow chart characteristics
CN114048294B (en) Similar population extension model training method, similar population extension method and device
CN110717787A (en) User classification method and device
CN116089708A (en) Agricultural knowledge recommendation method and device
Shankhdhar et al. Divorce prediction scale using improvised machine learning techniques
Lotfian et al. An approach for real-time validation of the location of biodiversity observations contributed in a citizen science project
Al-Kerboly et al. Clustering Algorithms Comparison for University of Anbar Researchers’ Google Scholar Profiles
CN113672746B (en) Multimedia resource recommendation method and device, electronic equipment and storage medium
CN113398569B (en) Card group classification processing, model training and card group searching method and equipment
CN114331789B (en) Intelligent cheap and clean knowledge recommendation method, device, equipment and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
REG Reference to a national code

Ref country code: HK

Ref legal event code: DE

Ref document number: 40050657

Country of ref document: HK

SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant