CN112711699B - User division method, system, computer device and readable storage medium - Google Patents

User division method, system, computer device and readable storage medium Download PDF

Info

Publication number
CN112711699B
CN112711699B CN201911016451.3A CN201911016451A CN112711699B CN 112711699 B CN112711699 B CN 112711699B CN 201911016451 A CN201911016451 A CN 201911016451A CN 112711699 B CN112711699 B CN 112711699B
Authority
CN
China
Prior art keywords
user
theme
manuscript data
users
label
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
CN201911016451.3A
Other languages
Chinese (zh)
Other versions
CN112711699A (en
Inventor
陈家伟
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.)
Shanghai Bilibili Technology Co Ltd
Original Assignee
Shanghai Bilibili Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Shanghai Bilibili Technology Co Ltd filed Critical Shanghai Bilibili Technology Co Ltd
Priority to CN201911016451.3A priority Critical patent/CN112711699B/en
Publication of CN112711699A publication Critical patent/CN112711699A/en
Application granted granted Critical
Publication of CN112711699B publication Critical patent/CN112711699B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/95Retrieval from the web
    • G06F16/953Querying, e.g. by the use of web search engines
    • G06F16/9535Search customisation based on user profiles and personalisation
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/95Retrieval from the web
    • G06F16/953Querying, e.g. by the use of web search engines
    • G06F16/9536Search customisation based on social or collaborative filtering
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00Pattern recognition
    • G06F18/20Analysing
    • G06F18/23Clustering techniques
    • YGENERAL 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
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02PCLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
    • Y02P90/00Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
    • Y02P90/30Computing systems specially adapted for manufacturing

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The invention discloses a user dividing method, a user dividing system, computer equipment and a readable storage medium, wherein a plurality of manuscript data uploaded by a first user are obtained, a theme corresponding to each manuscript data is generated according to the plurality of manuscript data, then operation information of a second user on the first user within preset time is obtained, a corresponding heterogeneous network is generated according to the theme and the operation information, and then the second user is clustered according to the heterogeneous network, so that the second user of the same type forms a community according to a clustering result, and the second user is divided. The invention can mine and process multi-type data to realize accurate division of users, and further provide more refined service for the users according to the division results of the users.

Description

User division method, system, computer device and readable storage medium
Technical Field
The invention relates to the field of big data, in particular to a user partitioning method, a user partitioning system, computer equipment and a readable storage medium.
Background
With the rapid development of the internet, the information amount of each large website is rapidly increased, and it becomes extremely difficult to classify users of each large website from massive data. At present, the general method for user division mainly comprises: (1) user segmentation based on demographic attributes; (2) user partitioning based on social relationships.
However, on one hand, due to the diversity of data types and the limitation of basic data accumulation, user partitioning is not combined with diverse data for effective mining, so that the partitioning result of the user is not accurate, and thus more refined services cannot be provided for the user. On the other hand, even if the diversity of data makes it possible to find the cause, higher demands are made on the mining method. Therefore, the invention aims to solve the problem of how to mine and process multi-type data to realize accurate division of users so as to provide more refined service for the users.
Disclosure of Invention
The invention aims to provide a user division method, a user division system, computer equipment and a readable storage medium, which can mine and process multi-type data to realize accurate division of users and further provide more refined services for the users according to the division results of the users.
According to an aspect of the present invention, there is provided a user segmentation method, including the steps of:
acquiring a plurality of manuscript data uploaded by a first user, and generating a theme corresponding to each manuscript data according to the plurality of manuscript data, wherein each manuscript data comprises a manuscript content and at least one label;
acquiring operation information of a second user on the first user within preset time;
generating a corresponding heterogeneous network according to the theme and the operation information;
and clustering the second users according to the heterogeneous network so that the second users of the same class form a community according to clustering results to divide the second users.
Optionally, the obtaining multiple pieces of manuscript data uploaded by the first user, and generating a theme corresponding to each piece of manuscript data according to the multiple pieces of manuscript data includes:
screening the labels of the manuscript data according to a preset rule to screen out a plurality of target labels;
acquiring the common occurrence times of the target labels in the manuscript data;
analyzing the distance of the target labels in the heterogeneous network according to the times, and training the target labels into corresponding target label word vectors;
and clustering the plurality of target labels according to the target label word vectors, the distance and the preset algorithm to form a theme corresponding to each manuscript data.
Optionally, the screening the tags of the multiple manuscript data according to a preset rule to screen out multiple target tags, including:
acquiring the labels of the plurality of manuscript data of the first user from a database, and counting the number of the manuscript contents under each label;
acquiring the heat information of each label from a database;
and screening the labels of the plurality of manuscript data according to the manuscript content quantity under each label and the heat information of each label so as to screen out the plurality of target labels.
Optionally, the generating a corresponding heterogeneous network according to the theme and the operation information includes:
calculating the distance between each label and each theme;
counting the average distance between the labels of the manuscript data and each topic according to the distance, and acquiring the score of the first user under each topic according to the average distance and a preset score mapping table;
and acquiring a first association relation between the first user and each topic according to the score.
Optionally, the counting average distances between the labels of the multiple pieces of manuscript data and the respective topics according to the distances to obtain scores of the first user under the respective topics according to the average distances and a preset score mapping table includes:
adding the distances from each label of the first user to each theme to obtain the total distance from the first user to each theme;
dividing the total distance by the number of the labels of the plurality of manuscript data to obtain an average distance between the first user and each theme;
and mapping the average distance to the score mapping table to obtain the score corresponding to the average distance.
Optionally, the generating a corresponding heterogeneous network according to the multiple topics and the operation information further includes:
counting the operation times of the second user on the first user according to the operation information;
acquiring total operation information executed by the second user within the preset time, and counting the total operation times of the second user within the preset time according to the total operation information;
calculating the weight relationship between the second user and the first user according to the operation times and the total operation times so as to obtain a second association relationship between the second user and the first user according to the weight relationship;
and generating the heterogeneous network according to the first incidence relation and the second incidence relation.
Optionally, the clustering the second users according to the heterogeneous network so that the second users of the same class form a community according to a clustering result includes:
acquiring a sampling range input by a third user;
taking the second user as a central node, sampling other second user nodes in the sampling range to obtain other second user node sequences in the sampling range;
training the other second user nodes according to the other second user node sequences to obtain corresponding other second user word vectors;
and clustering the second user and the other second users according to the word vectors of the other second users and the preset algorithm, and forming the community according to a clustering result.
In order to achieve the above object, the present invention further provides a user partitioning system, which specifically includes the following components:
the acquisition module is used for acquiring a plurality of manuscript data uploaded by a first user and operation information of a second user on the first user in preset time, wherein each manuscript data comprises manuscript contents and at least one label;
the generating module is used for generating a theme corresponding to each manuscript data according to the plurality of manuscript data and generating a corresponding heterogeneous network according to the theme and the operation information;
and the dividing module is used for clustering the second users according to the heterogeneous network so as to enable the second users of the same class to form a community according to a clustering result, and thus the second users are divided.
In order to achieve the above object, the present invention further provides a computer device, which specifically includes: a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the steps of the user segmentation method introduced above when executing the computer program.
To achieve the above object, the present invention also provides a computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, realizes the above-introduced steps of the user segmentation method.
The user dividing method, the user dividing system, the computer device and the readable storage medium provided by the invention are used for acquiring a plurality of manuscript data uploaded by a first user, generating a theme corresponding to each manuscript data according to the plurality of manuscript data, then acquiring operation information of a second user on the first user within preset time, generating a corresponding heterogeneous network according to the theme and the operation information, and then clustering the second users according to the heterogeneous network, so that the second users of the same type form a community according to a clustering result, and the second users are divided. The invention can mine and process the data of multiple types to realize the accurate division of the user, thereby providing more refined service for the user according to the division result of the user.
Drawings
Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are only for purposes of illustrating the preferred embodiments and are not to be construed as limiting the invention. Also, like reference numerals are used to refer to like parts throughout the drawings. In the drawings:
fig. 1 is an optional application environment diagram of the user partition method provided by the embodiment of the present disclosure;
fig. 2 is an alternative flow chart of a user partition method according to an embodiment of the present disclosure;
fig. 3 is a schematic diagram illustrating an alternative specific flowchart of step S100 in fig. 2;
fig. 4 is a schematic diagram of another alternative specific flowchart of step S104 in fig. 2;
fig. 5 is a schematic diagram illustrating an alternative specific flowchart of step S302 in fig. 4;
fig. 6 is a schematic diagram of another alternative specific flowchart of step S104 in fig. 2;
fig. 7 is a schematic diagram illustrating an alternative specific flowchart of step S106 in fig. 2;
FIG. 8 is a schematic diagram illustrating alternative program modules for a user segmentation system provided in an embodiment of the present disclosure;
fig. 9 is a schematic diagram of an alternative hardware architecture of a computer device according to an embodiment of the present disclosure.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. 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 invention.
FIG. 1 is a diagram of an application environment of the user partitioning method of the present invention. The method comprises the steps that a plurality of first users upload a plurality of manuscript data with labels to a video playing platform, and a user dividing system obtains the plurality of manuscript data and clusters the plurality of manuscript data into a plurality of themes according to the manuscript data. For example: the user dividing system receives manuscript data a1, a2 … and an, manuscript data b1, b2 … and bm, manuscript data k1, k2 … and kj uploaded by k first users, and then clusters the manuscript data a1, a2 … and an, the manuscript data b1, b2 … and bm, and the manuscript data k1, k2 … and kj to form a subject 1 and a subject 2. Exemplary clustering results are shown in fig. 1. Then, the user division system establishes the association relationship between the first user and each theme according to the distance between the manuscript data of the same first user and the theme. The second user realizes interaction with the first user by paying attention to the first user, commenting on the manuscript data of the first user, approving the manuscript data of the first user or collecting the manuscript data of the first user, so as to establish the incidence relation between the first user and the second user.
The following describes a user segmentation method provided by the present invention with reference to the accompanying drawings.
Fig. 2 is an optional flowchart of the user segmentation method of the present invention, and it is to be understood that the flowchart in this embodiment of the method is not used to limit the order of executing steps. The following description is made by taking a computer device as an execution subject. As shown in fig. 2, the method specifically includes steps S100 to S106.
Step S100: the method comprises the steps of obtaining a plurality of manuscript data uploaded by a first user, and generating a theme corresponding to each manuscript data according to the plurality of manuscript data, wherein each manuscript data comprises manuscript contents and at least one label.
Exemplarily, referring to fig. 1, when a first user 1 needs to upload a video about a food, tag remarking is performed on the video, for example, a food tag, and then the video with the food tag is uploaded, and a user classification system obtains the video with the food tag and generates a food theme according to the food tag and food tags of videos uploaded by other second users. When the first user 2 needs to upload the video about the exercise diet weight loss, the tags marking the video comprise food tags and sports tags, the video with the food tags and the sports tags is uploaded, the user dividing system obtains the video with the food tags and the sports tags, food themes are generated according to the food tags and other food tags, and sports themes are generated according to the sports tags and other sports tags uploaded by the second user. The manuscript content can also be an audio file or words.
Step S102: acquiring operation information of a second user to the first user within preset time, wherein the operation information comprises: paying attention to the first user, commenting on the manuscript data of the first user and/or collecting the manuscript data of the first user.
Specifically, after the first user uploads the video, the second user may perform an attention operation on the first user, or perform a comment operation, or a collection operation on the video uploaded by the first user. After the second user performs the operations, the user dividing system receives operation information of the first user by the second user. In this embodiment, in order to better explain the association relationship between the first user and the second user, that is, the interest level of the second user in the first user, the second user reviews, approves and collects the video uploaded by the first user, and the brief description is that the second user reviews, approves and collects the first user.
Step S104: and generating a corresponding heterogeneous network according to the theme and the operation information, wherein the heterogeneous network comprises a plurality of nodes, and the plurality of nodes comprise a first user node, a second user node and a theme node.
Specifically, after a theme is generated according to a label of manuscript data uploaded by a user, an association relation table of the first user, the second user and the theme is generated according to the theme and operation information of the second user on the first user, and the heterogeneous network is generated according to the association relation table. The more frequent the second user operates the first user, the higher the association degree between the second user and the first user is, and the shorter the distance between the second user and the first user is, the shorter the side length between the second user and the first user in the heterogeneous network is; on the contrary, the lower the association degree between the second user and the first user is, the longer the distance between the second user and the first user is, and the longer the side length between the second user and the first user in the heterogeneous network is.
Illustratively, as the videos uploaded by the first users have a plurality of tags, a plurality of topics are generated according to the tags, so as to establish the relationship between the first users and the topics. And generating a relationship between the first user and the second user according to the operation of the second user on the first user, further establishing an association relationship among the first user, the second user and the theme, and constructing a heterogeneous network taking the first user, the second user and the theme as nodes. It should be noted that the heterogeneous network is used to indicate the degree of association between the first user, the second user, and the theme. The shorter the side length between the first user and the second user is, the higher the interest degree of the second user in the first user is, and the higher the association degree between the second user and the first user is. The shorter the side length between the first user and the theme is, the closer the manuscript data uploaded by the first user is to the theme is, and the higher the association degree between the first user and the theme is.
With reference to fig. 1, a second user a performs a related operation on a first user 1, a first user 2 and a first user n to establish an association relationship between the second user a and the first user 1, the first user 2 and the first user n, a second user B performs a related operation on the first user 1 and the first user 2 to establish an association relationship between the second user B and the first user 1 and the first user 2, and a second user C performs a related operation on the first user 2 and the first user n to establish an association relationship between the second user C and the first user 2 and the first user n. The user dividing system generates a theme 1 according to the manuscript data a1 and a manuscript data a2 of the first user 1, the manuscript data b1 of the first user 2, the manuscript data b2 and a manuscript data b3, and the manuscript data k1 of the first user n, and generates a theme 2 according to the manuscript data a3 of the first user 1, the manuscript data k2 of the first user n and the manuscript data k3, so as to establish an association relationship between the first user 1, the first user 2, and the first user n and the theme 1 and the theme 2. And the user dividing system generates a heterogeneous network according to the association relationship between the second user A and the first user 1, the association relationship between the second user B and the first user 1 and the first user 2, the association relationship between the second user C and the first user 2 and the association relationship between the second user C and the first user n and the association relationship between the first user 1 and the first user 2, and the association relationship between the first user n and the subject 1 and the subject 2.
Step S106: and clustering the second users according to the heterogeneous network so that the second users of the same class form a community according to clustering results to divide the second users.
Illustratively, second users having similar interests according to the heterogeneous networks are divided into the same community to complete the division of the second users. With reference to fig. 1, if the association relationship between a second user a and a first user 1 of all first users is the closest, and the association relationship between a second user B and the first user 1 of all first users is the closest, that is, the second user a is most interested in the first user 1, the second user B is also most interested in the first user 1, and the first user 1 is the closest to the subject 1, the second user a and the second user B are grouped into one group, the second user a and the second user B form a community, and the second user a and the second user B are further divided into the same community.
In an exemplary embodiment, as shown in fig. 3, the step S100 may include steps S200 to S206.
Step S200: and screening the labels of the manuscript data according to a preset rule so as to screen out a plurality of target labels.
In a preferred embodiment, the step of filtering the tags of the manuscript data according to a preset rule to filter out a plurality of target tags includes: acquiring labels of a plurality of manuscript data of the first user and the heat information of each label from a database, counting the number of the manuscript contents under each label, and screening the labels of the plurality of manuscript data according to the number of the manuscript contents under each label and the heat information of each label to screen a plurality of target labels from the plurality of labels.
For example, referring to fig. 1, since the number of videos uploaded by the first user 1 and the first user 2 … is large, and each video has at least one tag, by acquiring all tags of all videos uploaded by the first users 1 to the first user n and popularity information (i.e. popularity information) of each tag, and counting the number of videos under each tag, for example: only the label of the manuscript data a1 is food, and the label of the manuscript data a2 is food and travel, so that the number of videos under the food label is 2. And counting the quality of each label by integrating the number of videos under each label and the popularity information of each label so as to screen out target labels (namely high-quality labels) with a large number of videos and high popularity from all the labels. Of course, the number of the screened target tags can be set arbitrarily according to the situation.
Step S202: and acquiring the common occurrence times of the target labels in the manuscript data.
Step S204: and analyzing the distance of the target labels in the heterogeneous network according to the times, and training the target labels into corresponding target label word vectors. It should be noted that, the more times two target tags appear together in multiple manuscript data, the closer the two target tags are in the heterogeneous network.
For example, for weight loss, it is often necessary to exercise while controlling diet. Therefore, two kinds of labels, namely a food label and a motion label, often appear in a video about weight loss, and the times of the common appearance of the food label and the motion label in manuscript data are obtained, so that the distance degree of the food label and the motion label in a heterogeneous network is analyzed according to the times of the common appearance of the food label and the motion label in the manuscript data. And then, training the food labels and the motion labels into corresponding food label word vectors and motion label word vectors.
Step S206: and clustering the plurality of target labels according to the target label word vectors, the distance and the preset algorithm to form a theme corresponding to each manuscript data. The preset algorithm may include any one of a Kmeans clustering algorithm, a mean shift clustering algorithm, and a density-based clustering algorithm.
Illustratively, in combination with the degree of distance of the food labels and the motion labels in the heterogeneous network, kmeans clustering is adopted to cluster the trained food label word vectors and motion label word vectors to form the same topic, for example: subject of weight loss. For another example, the educational label and the travel label have a small number of common occurrences in the same manuscript data and a long distance in a heterogeneous network, so that word vectors generated by the educational label and the travel label form different topics, for example: respectively an educational theme and a tourist theme. Through the exemplary embodiment, the corresponding theme can be generated according to the co-occurrence frequency of the label under the manuscript, so that the incidence relation between the label and the theme is established.
In an exemplary embodiment, as shown in fig. 4, the step S104 may include steps S300 to S304.
Step S300: and calculating the distance between each label and each theme.
Illustratively, the distance function in the Kmeans algorithm is adopted to calculate the distance between each label and each topic. For example, in conjunction with fig. 1, fig. 1 illustrates only two topics 1 and 2 as examples. Calculating the distance from the food label in the manuscript data a1 to the center of the theme 1 and the center of the theme 2 respectively, calculating the distance from the food label and the travel label in the manuscript data a2 to the center of the theme 1 and the center of the theme 2 respectively, and sequentially calculating the distance from the label in all the manuscript data to the center of the theme 1 and the center of the theme 2 respectively.
Step S302: and counting the average distance between the labels of the manuscript data and each theme according to the distance so as to obtain the scores of the first user under each theme according to the average distance and a preset score mapping table.
Exemplarily, referring to fig. 1, a first user 1 includes manuscript data a1 and manuscript data a2 … manuscript data an, and obtains a score of the first user 1 under the theme 1 by calculating an average distance from all labels in the manuscript data an to the center of the theme 1 in the manuscript data a1 and the manuscript data a2 …, and obtaining a score corresponding to the average distance according to a preset score mapping table. It should be noted that the higher the average distance, the higher the score of the first user 1. Of course, other rules of the score mapping table may be set, for example, the higher the average distance, the lower the score of the first user. In the embodiment of the present invention, the higher the average distance is, the higher the score of the first user is taken as an example for explanation.
Step S304: and acquiring a first association relation between the first user and each topic according to the score.
Illustratively, the higher the score of the first user 1 under the topic 1, the more closely the first user 1 has an association relationship with the topic 1. By the exemplary embodiment, the distances from all the labels of the manuscript data of the first user to all the topics can be integrated, and the association relationship between the first user and each topic can be accurately obtained.
In an exemplary embodiment, as shown in fig. 5, the step S302 may include steps S400 to S404.
Step S400: and adding the distances from each label of the first user to each theme to obtain the total distance from the first user to each theme.
Exemplarily, in conjunction with fig. 1, the distances from all the labels in the manuscript data a1 and a2 … of the first user 1 and the manuscript data a3 to the center of the theme 1 are added to obtain the total distance from the first user 1 to the center of the theme 1. Adding the distances from all the labels in the manuscript data a1 and a2 … of the first user 1 to the center of the theme 2 to obtain the total distance from the first user 1 to the center of the theme 2. The total distance from the first user 2 to the first user n to the subject 1 and the subject 2 is calculated accordingly, which is not described herein again.
Step S402: and dividing the total distance by the number of the labels of the plurality of manuscript data to obtain the average distance between the first user and each theme.
Illustratively, the number of all the labels in all the manuscript data of the first user 1 is counted, and the calculated total distance from the first user 1 to the center of the theme 1 is divided by the number of all the labels, so as to obtain the average distance from the first user 1 to the center of the theme 1. And setting the calculated total distance from the first user 1 to the center of the theme 2 in the number of all the labels to obtain the average distance from the first user 1 to the center of the theme 2. The average distance from the first user 2 to the first user n to the subject 1 and the subject 2 is calculated accordingly, which is not described herein again.
Step S404: and mapping the average distance to the score mapping table to obtain the score corresponding to the average distance.
Illustratively, the calculated average distance from the first user 1 to the subject 1 is mapped with the preset score mapping table, and a score corresponding to the average distance is obtained from the score mapping table according to a mapping result.
In an exemplary embodiment, as shown in fig. 6, the step S104 may include steps S500 to S506.
Step S500: and counting the operation times of the second user on the first user according to the operation information.
Exemplarily, referring to fig. 1, operation information of attention, praise, comment and collection of a second user a to the first user 1 is obtained, and the number of operations of the second user a to the first user 1 is counted according to the operation information.
Step S502: and acquiring total operation information executed by the second user within the preset time, and counting the total operation times of the second user within the preset time according to the total operation information.
Exemplarily, with reference to fig. 1, the total operation information executed by the second user a in the preset time is obtained, which includes the operation information of the second user a on the first user 2 to the first user n, and the total operation times of the second user a in the preset time is counted according to the total operation information.
Step S504: and calculating the weight relationship between the second user and the first user according to the operation times and the total operation times so as to obtain a second association relationship between the second user and the first user according to the weight relationship.
For example, the operation frequency of the second user a on the first user 1 is divided by the total operation frequency of the second user a in the preset time to obtain a weight of the first user 1 in the second user a, so as to obtain an association relationship between the second user a and the first user 1 according to the weight, that is, the interest degree of the second user a on the first user 1. The association relationship between the second user a and the first users 2 to N and the association relationship between the first users B and C and the first users 2 to N are calculated accordingly, which is not described herein again.
In a preferred embodiment, the user partitioning system further presets a weight distance table, and when obtaining the weight of the first user 1 occupying the second user a, matches the weight with a preset distance in the weight distance table to obtain a distance corresponding to the weight, so as to obtain the distance between the second user a and the first user 1 in the heterogeneous network.
Step S506: and generating the heterogeneous network according to the first incidence relation and the second incidence relation. Illustratively, the heterogeneous network is generated by combining the first association with the second association.
Through the exemplary embodiment, the association relationship between the second user and the first user can be established according to the operation information of the second user on the first user, and a network architecture basis is provided for the division of the second user by combining the association relationship heterogeneous network between the first user and each theme.
In an exemplary embodiment, as shown in fig. 7, the step S106 may include steps S600 to S606.
Step S600: a sampling range of a third user input is obtained.
Step S602: and taking the second user as a central node, and sampling other second user nodes in the sampling range to obtain other second user node sequences in the sampling range.
Referring to fig. 1, a sampling distance of a third user is obtained, and other second user nodes (for example, a second user B and a second user C within the sampling distance) on a node in the heterogeneous network are sampled by using a second user a as a center node. Wherein the other second user nodes include a second user B who has a co-operation with the second user a (e.g. fig. 1, a first user 1 and a first user 2 who are co-operated with the second user a) and the second user C who has a same topic of interest as the second user a.
Step S604: and training the other second user nodes according to the other second user node sequences to obtain corresponding other second user word vectors.
Step S606: and clustering the second user and the other second users according to the word vectors of the other second users and the preset algorithm, and forming the community according to a clustering result.
Illustratively, the second user a, the second user B and the second user C are clustered to form a community by using a Kmeans algorithm. According to the exemplary embodiment, the second users in the preset sampling range are sampled to cluster the second users in the sampling range, so that the second users can be divided more finely.
Based on the user partition method provided in the foregoing embodiment, the present embodiment provides a user partition system, and the user partition system may be applied to a computer device. In particular, FIG. 8 illustrates an alternative block diagram of the user-partitioned system, which is partitioned into one or more program modules, which are stored in a storage medium and executed by one or more processors to implement the present invention. The program module referred to in the present invention is a series of computer program instruction segments capable of performing specific functions, and is more suitable for describing the execution process of the user partitioning system in the storage medium than the program itself, and the following description will specifically describe the functions of each program module in this embodiment.
As shown in fig. 8, the user segmentation system specifically includes the following components:
an obtaining module 201, configured to obtain multiple pieces of manuscript data uploaded by a first user and operation information of a second user on the first user within a preset time, where each piece of manuscript data includes manuscript content and at least one tag, and the operation information includes: paying attention to the first user, commenting on the manuscript data of the first user and collecting the manuscript data of the first user.
Illustratively, in conjunction with fig. 1, when the first user 1 needs to upload a video about a food, tag remarks are made to the video, such as a food tag, and then the video with the food tag is uploaded, and the obtaining module 201 obtains the video with the food tag. When the first user 2 needs to upload a video about diet slimming, the tags labeling the video include a food tag and a sports tag, and upload the video with the food tag and the sports tag, and the acquiring module 201 acquires the video with the food tag and the sports tag. The manuscript content can also be an audio file or words.
Specifically, after the first user uploads the video, the second user may perform an attention operation on the first user, or perform a comment operation, or a collection operation on the video uploaded by the first user. After the second user performs the operations, the user dividing system receives operation information of the first user by the second user. In this embodiment, in order to better explain the association relationship between the first user and the second user, that is, the interest level of the second user in the first user, the second user reviews, approves and collects the video uploaded by the first user, and the brief description is that the second user reviews, approves and collects the first user.
A generating module 202, configured to generate a theme corresponding to each piece of manuscript data according to the plurality of pieces of manuscript data, and generate a corresponding heterogeneous network according to the theme and the operation information, where the heterogeneous network includes a plurality of nodes, and the plurality of nodes includes a first user node, a second user node, and a theme node.
Specifically, after a theme is generated according to a label of manuscript data uploaded by a user, an association relation table of the first user, the second user and the theme is generated according to the theme and operation information of the second user on the first user, and the heterogeneous network is generated according to the association relation table. The more frequent the second user operates the first user, the higher the association degree between the second user and the first user is, and the closer the distance between the second user and the first user is, the shorter the side length between the second user and the first user in the heterogeneous network is; conversely, the lower the degree of association between the second user and the first user is, the longer the distance between the second user and the first user is, and the longer the side length between the second user and the first user in the heterogeneous network is.
Illustratively, since there are multiple tags in the videos uploaded by multiple first users, the generating module 202 generates multiple topics according to the multiple tags, so as to establish the relationship between the first users and the respective topics. For example: generating a food theme according to the food label of the first user 1 and food labels of videos uploaded by other first users (such as the first user 2 and the first user 3); and generating a sports theme according to the sports label of the first user and the sports labels uploaded by other first users. And then, according to the operation of a second user on the first user, generating the relationship between the first user and the second user, further establishing an association relationship among the first user, the second user and the theme, and constructing a heterogeneous network taking the first user, the second user and the theme as nodes. It should be noted that the heterogeneous network is used to indicate the degree of association between the first user, the second user and the theme. The shorter the side length between the first user and the second user is, the higher the interest degree of the second user in the first user is, and the higher the association degree between the second user and the first user is. The shorter the side length between the first user and the theme is, the closer the manuscript data uploaded by the first user is to the theme is, and the higher the association degree between the first user and the theme is.
With reference to fig. 1, a second user a performs a related operation on a first user 1, a first user 2 and a first user n to establish an association relationship between the second user a and the first user 1, between the first user 2 and the first user n, a second user B performs a related operation on the first user 1 and the first user 2 to establish an association relationship between the second user B and the first user 1 and between the second user B and the first user 2, and a second user C performs a related operation on the first user 2 and the first user n to establish an association relationship between the second user C and the first user 2 and between the first user n. The generating module 202 generates a theme 1 according to the manuscript data a1 and a2 of the first user 1, the manuscript data b1 of the first user 2, the manuscript data b2 and b3, and the manuscript data k1 of the first user n, and generates a theme 2 according to the manuscript data a3 of the first user 1, the manuscript data k2 and the manuscript data k3 of the first user n, so as to establish an association relationship between the first user 1, the first user 2, the first user n, the theme 1, and the theme 2. The generating module 202 generates a heterogeneous network according to the association relationship between the second user a and the first user 1, the association relationship between the first user 2 and the first user n, the association relationship between the second user B and the first user 1 and the first user 2, and the association relationship between the second user C and the first user 2 and the first user n and the association relationship between the first user 1, the first user 2, and the association relationship between the first user n and the subject 1 or the subject 2.
The dividing module 203 is configured to cluster the second users according to the heterogeneous network, so that the second users of the same class form a community according to a clustering result, and divide the second users.
Illustratively, the partitioning module 203 partitions second users with similar interests in the heterogeneous network into a same community to complete the partitioning of the second users. With reference to fig. 1, if the association relationship between a second user a and a first user 1 of all first users is the closest, and the association relationship between a second user B and the first user 1 of all first users is the closest, that is, the second user a is most interested in the first user 1, the second user B is also most interested in the first user 1, and the first user 1 is the closest to the subject 1, the dividing module 203 groups the second user a and the second user B into one group, the second user a and the second user B form a community, and further divides the second user a and the second user B into the same community.
In an exemplary embodiment, the generating module 202 further includes a screening unit, an analyzing unit, and a first clustering unit:
and the screening unit is used for screening the labels of the manuscript data according to a preset rule so as to screen out a plurality of target labels.
In a preferred embodiment, the screening unit is further configured to obtain tags of the multiple pieces of manuscript data of the first user and heat information of each tag from a database, count the number of contents of the manuscript under each tag, and then screen the tags of the multiple pieces of manuscript data according to the number of contents of the manuscript under each tag and the heat information of each tag, so as to screen multiple target tags from the multiple tags.
For example, with reference to fig. 1, because the number of videos uploaded by the first user 1 and the first user 2 … by the first user n is large, and each video has at least one tag, by obtaining all tags of all videos uploaded by the first users 1 to the first user n and popularity information (i.e., popularity information) of each tag, and counting the number of videos under each tag, for example: only the label of the manuscript data a1 is food, the label of the manuscript data a2 is food and travel, and the number of videos under the food label is 2. By integrating the number of videos under each label and the popularity information of each label, the quality of each label is counted, so that the screening unit screens out target labels (namely, high-quality labels) with a large number of videos and high popularity from all the labels. Of course, the number of the screened target tags can be set arbitrarily according to the situation.
The first acquisition unit is configured to acquire the number of times that the plurality of target tags commonly appear in the plurality of manuscript data.
And the analysis unit is used for analyzing the distance of the target labels in the heterogeneous network according to the times and training the target labels into corresponding target label word vectors. It should be noted that, the more times two target tags appear together in multiple manuscript data, the closer the two target tags are in the heterogeneous network.
For example, for weight loss, it is often necessary to exercise while controlling diet. Therefore, in a video about weight loss, two tags, namely a food tag and a motion tag, often appear, and at this time, the first obtaining unit obtains the times of the common appearance of the food tag and the motion tag in the manuscript data, so that the analyzing unit analyzes the distance between the food tag and the motion tag in the heterogeneous network according to the times of the common appearance of the food tag and the motion tag in the manuscript data. And then, training the food labels and the motion labels into corresponding food label word vectors and motion label word vectors.
The first clustering unit is used for clustering the plurality of target labels according to the target label word vectors, the distance and the preset algorithm so as to form a theme corresponding to each manuscript data. The preset algorithm may include any one of a Kmeans clustering algorithm, a mean shift clustering algorithm, and a density-based clustering algorithm.
Illustratively, in combination with the degree of distance of the food labels and the motion labels in the heterogeneous network, kmeans clustering is adopted to cluster the trained food label word vectors and motion label word vectors to form the same topic, for example: subject of weight loss. For another example, the educational label and the travel label have a small number of common occurrences in the same manuscript data and a long distance in a heterogeneous network, so that word vectors generated by the educational label and the travel label form different topics, for example: respectively an educational theme and a tourist theme. Through the exemplary embodiment, the corresponding theme can be generated according to the co-occurrence frequency of the label under the manuscript, so that the incidence relation between the label and the theme is established.
In an exemplary embodiment, the generating module 202 further comprises a computing unit.
The calculating unit is used for calculating the distance between each label and each theme.
Illustratively, the calculation unit calculates the distance between each label and each topic by using a distance function in a Kmeans algorithm. For example, in conjunction with fig. 1, fig. 1 illustrates only two topics 1 and 2 as examples. The calculating unit calculates the distances from the food labels in the manuscript data a1 to the centers of the theme 1 and the theme 2 respectively, calculates the distances from the food labels and the travel labels in the manuscript data a2 to the centers of the theme 1 and the theme 2 respectively, and calculates the distances from the labels in all the manuscript data to the centers of the theme 1 and the theme 2 respectively in sequence.
The calculating unit is further configured to count average distances between the labels of the multiple pieces of manuscript data and the respective topics according to the distances, so as to obtain scores of the first user under the respective topics according to the average distances and a preset score mapping table.
Exemplarily, referring to fig. 1, the first user 1 includes manuscript data a1 and manuscript data a2 … manuscript data an, and the calculating unit calculates an average distance from all tags in the manuscript data a1, the manuscript data a2 … and the manuscript data an to the center of the topic 1, and obtains a score corresponding to the average distance according to a preset score mapping table, so as to obtain the score of the first user 1 under the topic 1. It should be noted that the higher the average distance, the higher the score of the first user 1. Of course, other rules of the score mapping table may be set, for example, the higher the average distance, the lower the score of the first user. In the embodiment of the present invention, the higher the average distance is, the higher the score of the first user is taken as an example for explanation.
The first obtaining unit is used for obtaining a first association relation between the first user and each topic according to the score.
Illustratively, the higher the score of the first user 1 under the topic 1, the more closely the first user 1 has an association relationship with the topic 1. By the exemplary embodiment, the distances from all the labels of the manuscript data of the first user to all the topics can be integrated, and the association relationship between the first user and each topic can be accurately obtained.
In an exemplary embodiment, the computing unit is further configured to:
adding the distances from each label of the first user to each theme to obtain the total distance from the first user to each theme; dividing the total distance by the number of the labels of the plurality of manuscript data to obtain an average distance between the first user and each theme; and mapping the average distance to the score mapping table to obtain the score corresponding to the average distance.
Exemplarily, with reference to fig. 1, first, the calculation unit adds distances from all tags in the manuscript data a1 and a manuscript data a2 … of the first user 1 to the center of the theme 1 to obtain a total distance from the first user 1 to the center of the theme 1. Adding the distances from all the labels in the manuscript data a1 and a2 … of the first user 1 to the center of the theme 2 to obtain the total distance from the first user 1 to the center of the theme 2. The total distance from the first user 2 to the first user n to the theme 1 and the theme 2 is calculated accordingly, which is not described herein again.
Then, counting the number of all the labels in all the manuscript data of the first user 1, and dividing the calculated total distance from the first user 1 to the center of the theme 1 by the number of all the labels to obtain the average distance from the first user 1 to the center of the theme 1. And setting the calculated total distance from the first user 1 to the center of the theme 2 in the number of all the labels to obtain the average distance from the first user 1 to the center of the theme 2. The average distance from the first user 2 to the first user n to the subject 1 and the subject 2 is calculated accordingly, which is not described herein again.
And finally, mapping the calculated average distance from the first user 1 to the subject 1 with the preset score mapping table, and acquiring a score corresponding to the average distance from the score mapping table according to a mapping result.
In an exemplary embodiment, the generating module 202 further includes a statistic unit and a generating unit.
And the counting unit is used for counting the operation times of the second user on the first user according to the operation information.
For example, referring to fig. 1, after the obtaining module 201 obtains operation information of attention, praise, comment and collection of a second user a to the first user 1, the counting unit counts the operation times of the second user a to the first user 1 according to the operation information.
The first obtaining unit is further configured to obtain total operation information executed by the second user within the preset time.
The counting unit is further configured to count the total operation times of the second user in the preset time according to the total operation information.
Illustratively, with reference to fig. 1, the first obtaining unit obtains total operation information, which is executed by the second user a within a preset time and includes operation information of the second user a on the first user 2 to the first user n, and then the counting unit counts the total operation times of the second user a within the preset time according to the total operation information.
The calculating unit is configured to calculate a weight relationship between the second user and the first user according to the operation times and the total operation times, so as to obtain a second association relationship between the second user and the first user according to the weight relationship.
For example, the operation frequency of the second user a on the first user 1 is divided by the total operation frequency of the second user a in the preset time to obtain a weight of the first user 1 in the second user a, so as to obtain an association relationship between the second user a and the first user 1 according to the weight, that is, the interest degree of the second user a on the first user 1. The association relationship between the second user a and the first users 2 to N and the association relationship between the first users B and C and the first users 2 to N are calculated accordingly, which is not described herein again.
In a preferred embodiment, the user partitioning system further presets a weight distance table, and when obtaining the weight of the first user 1 occupying the second user a, matches the weight with a preset distance in the weight distance table to obtain a distance corresponding to the weight, so as to obtain the distance between the second user a and the first user 1 in the heterogeneous network.
The generating unit is configured to generate the heterogeneous network according to the first association relationship and the second association relationship. Illustratively, the heterogeneous network is generated by combining the first association with the second association. Through the exemplary embodiment, the association relationship between the second user and the first user can be established according to the operation information of the second user on the first user, and a network architecture basis is provided for the division of the second user by combining the association relationship heterogeneous network between the first user and each theme.
In an exemplary embodiment, the partitioning module 203 includes a second obtaining unit, a sampling unit, a training unit, and a second clustering unit.
The second acquisition unit is used for acquiring a sampling range input by a third user.
And the sampling unit is used for sampling other second user nodes in the sampling range by taking the second user as a central node so as to obtain other second user node sequences in the sampling range.
Referring to fig. 1, the second obtaining unit obtains a sampling distance of a third user, and then the sampling unit samples other second user nodes (for example, a second user B and a second user C within the sampling distance) on a node in the heterogeneous network by using a second user a as a center node. Wherein the other second user nodes include a second user B having a common operation with the second user a (as in fig. 1, a first user 1 and a first user 2 of the second user B and the second user a) and a second user C having a same topic of interest with the second user a.
And the training unit is used for training the other second user nodes according to the other second user node sequences so as to obtain corresponding other second user word vectors.
The second clustering unit is used for clustering the second user and the other second users according to the word vectors of the other second users and the preset algorithm, and forming the community according to a clustering result.
Illustratively, the second user a, the second user B and the second user C are clustered to form a community by using a Kmeans algorithm. According to the exemplary embodiment, the second users in the preset sampling range are sampled to cluster the second users in the sampling range, so that the second users can be divided more finely.
The embodiment also provides a computer device, such as a smart phone, a tablet computer, a notebook computer, a desktop computer, a rack server, a blade server, a tower server or a rack server (including an independent server or a server cluster composed of a plurality of servers) capable of executing programs, and the like. As shown in fig. 9, the computer device 30 of the present embodiment includes at least, but is not limited to: a memory 301, a processor 302 communicatively coupled to each other via a system bus. It is noted that fig. 9 only shows a computer device 30 with components 301-302, but it is to be understood that not all shown components are required to be implemented, and that more or fewer components may be implemented instead.
In this embodiment, the memory 301 (i.e., the readable storage medium) includes a flash memory, a hard disk, a multimedia card, a card-type memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, and the like. In some embodiments, the storage 301 may be an internal storage unit of the computer device 30, such as a hard disk or a memory of the computer device 30. In other embodiments, the memory 301 may also be an external storage device of the computer device 30, such as a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), or the like, provided on the computer device 30. Of course, the memory 301 may also include both internal and external storage devices for the computer device 30. In the present embodiment, the memory 301 is generally used for storing an operating system and various types of application software installed on the computer device 30, such as the program codes of the user partition system of the above-described embodiment. In addition, the memory 301 may also be used to temporarily store various types of data that have been output or are to be output.
Processor 302 may be a Central Processing Unit (CPU), controller, microcontroller, microprocessor, or other data Processing chip in some embodiments. The processor 302 generally serves to control the overall operation of the computer device 30.
Specifically, in this embodiment, the processor 302 is configured to execute the program of the user partition method stored in the processor 302, and when executed, the program of the user partition method implements the following steps:
acquiring a plurality of manuscript data uploaded by a first user, and generating a theme corresponding to each manuscript data according to the plurality of manuscript data, wherein each manuscript data comprises manuscript contents and at least one label;
acquiring operation information of a second user on the first user within preset time;
generating a corresponding heterogeneous network according to the theme and the operation information;
and clustering the second users according to the heterogeneous network so that the second users of the same class form a community according to clustering results to divide the second users.
For the specific embodiment of the process of the above method steps, reference may be made to the above embodiments, and details of this embodiment are not repeated herein.
The present embodiments also provide a computer readable storage medium, such as a flash memory, a hard disk, a multimedia card, a card type memory (e.g., SD or DX memory, etc.), a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a read-only memory (ROM), an electrically erasable programmable read-only memory (EEPROM), a programmable read-only memory (PROM), a magnetic memory, a magnetic disk, an optical disk, a server, an App, etc., having stored thereon a computer program that when executed by a processor implements the method steps of:
acquiring a plurality of manuscript data uploaded by a first user, and generating a theme corresponding to each manuscript data according to the plurality of manuscript data, wherein each manuscript data comprises manuscript contents and at least one label;
acquiring operation information of a second user to the first user within preset time;
generating a corresponding heterogeneous network according to the theme and the operation information;
and clustering the second users according to the heterogeneous network so that the second users of the same class form a community according to clustering results to divide the second users.
For the specific embodiment of the process of the above method steps, reference may be made to the above embodiments, and details are not repeated here.
The user partitioning method, the user partitioning system, the computer device, and the readable storage medium provided in this embodiment acquire multiple pieces of manuscript data uploaded by a first user, generate a theme corresponding to each piece of manuscript data according to the multiple pieces of manuscript data, then acquire operation information of a second user on the first user within a preset time, generate a corresponding heterogeneous network according to the theme and the operation information, and then cluster the second users according to the heterogeneous network, so that the second users of the same class form a community according to a clustering result to partition the second users. The invention can mine and process multi-type data to realize accurate division of users, and further provide more refined service for the users according to the division results of the users.
It should be noted that, in this document, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising a … …" does not exclude the presence of another identical element in a process, method, article, or apparatus that comprises the element.
The above-mentioned serial numbers of the embodiments of the present invention are merely for description and do not represent the merits of the embodiments.
Through the description of the foregoing embodiments, it is clear to those skilled in the art that the method of the foregoing embodiments may be implemented by software plus a necessary general hardware platform, and certainly may also be implemented by hardware, but in many cases, the former is a better implementation.
The above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention, and all equivalent structures or equivalent processes performed by the present invention or directly or indirectly applied to other related technical fields are also included in the scope of the present invention.

Claims (10)

1. A method for user segmentation, the method comprising:
acquiring a plurality of manuscript data uploaded by a first user, and generating a theme corresponding to each manuscript data according to the plurality of manuscript data, wherein each manuscript data comprises a manuscript content and at least one label;
acquiring operation information of a second user on the first user within preset time;
generating a corresponding heterogeneous network according to the theme and the operation information, wherein the heterogeneous network comprises a first user node, a second user node and a theme node, and the side length between the first user node and the second user node and the side length between the first user node and the theme node represent the association degree among the first user, the second user and the theme;
and clustering the second users according to the heterogeneous network so that the second users of the same class form a community according to clustering results to divide the second users.
2. The user classifying method according to claim 1, wherein the obtaining of a plurality of manuscript data uploaded by a first user and the generating of a theme corresponding to each of the manuscript data according to the plurality of manuscript data comprises:
screening the labels of the manuscript data according to a preset rule to screen out a plurality of target labels;
acquiring the common occurrence times of the target labels in the manuscript data;
analyzing the distance of the target labels in the heterogeneous network according to the times, and training the target labels into corresponding target label word vectors;
and clustering the plurality of target labels according to the target label word vectors, the distance and the preset algorithm to form a theme corresponding to each manuscript data.
3. The user segmentation method according to claim 2, wherein the filtering the tags of the plurality of manuscript data according to a preset rule to filter out a plurality of target tags comprises:
acquiring the labels of the plurality of manuscript data of the first user from a database, and counting the number of the manuscript contents under each label;
acquiring the heat information of each label from a database;
and screening the labels of the plurality of manuscript data according to the manuscript content quantity under each label and the heat information of each label so as to screen out the plurality of target labels.
4. The user partitioning method according to claim 3, wherein the generating of the corresponding heterogeneous network according to the theme and the operation information includes:
calculating the distance from each label to each theme;
counting the average distance between the labels of the manuscript data and each topic according to the distance, and acquiring the score of the first user under each topic according to the average distance and a preset score mapping table;
and acquiring a first association relation between the first user and each topic according to the score.
5. The user segmentation method according to claim 4, wherein the calculating an average distance between the labels of the plurality of manuscript data and each topic according to the distance to obtain the score of the first user under each topic according to the average distance and a preset score mapping table comprises:
adding the distances from each label of the first user to each theme to obtain the total distance from the first user to each theme;
dividing the total distance by the number of the labels of the plurality of manuscript data to obtain an average distance between the first user and each theme;
and mapping the average distance to the score mapping table to obtain the score corresponding to the average distance.
6. The user partitioning method according to claim 4, wherein the generating of the corresponding heterogeneous network according to the theme and the operation information further comprises:
counting the operation times of the second user on the first user according to the operation information;
acquiring total operation information executed by the second user within the preset time, and counting the total operation times of the second user within the preset time according to the total operation information;
calculating the weight relationship between the second user and the first user according to the operation times and the total operation times so as to obtain a second association relationship between the second user and the first user according to the weight relationship;
and generating the heterogeneous network according to the first incidence relation and the second incidence relation.
7. The method as claimed in claim 1, wherein the clustering the second users according to the heterogeneous network so that the second users of the same class form a community according to the clustering result comprises:
acquiring a sampling range input by a third user;
taking the second user as a central node, sampling other second user nodes in the sampling range to obtain other second user node sequences in the sampling range;
training the other second user nodes according to the other second user node sequences to obtain corresponding other second user word vectors;
and clustering the second user and the other second users according to the word vectors of the other second users and the preset algorithm, and forming the community according to a clustering result.
8. A user segmentation system, the system comprising:
the acquisition module is used for acquiring a plurality of manuscript data uploaded by a first user and operation information of a second user on the first user in preset time, wherein each manuscript data comprises manuscript contents and at least one label;
the generating module is used for generating a theme corresponding to each manuscript data according to the plurality of manuscript data and generating a corresponding heterogeneous network according to the theme and the operation information, wherein the heterogeneous network comprises a first user node, a second user node and a theme node, and the side length between the first user node and the second user node and the side length between the first user node and the theme node represent the association degree among the first user, the second user and the theme;
and the dividing module is used for clustering the second users according to the heterogeneous network so as to enable the second users of the same class to form a community according to a clustering result, and thus the second users are divided.
9. A computer device, the computer device comprising: memory, processor and computer program stored on the memory and executable on the processor, characterized in that the processor implements the steps of the user segmentation method of any one of claims 1 to 7 when executing the computer program.
10. A computer-readable storage medium, on which a computer program is stored, which computer program, when being executed by a processor, carries out the steps of the user segmentation method according to any one of claims 1 to 7.
CN201911016451.3A 2019-10-24 2019-10-24 User division method, system, computer device and readable storage medium Active CN112711699B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911016451.3A CN112711699B (en) 2019-10-24 2019-10-24 User division method, system, computer device and readable storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911016451.3A CN112711699B (en) 2019-10-24 2019-10-24 User division method, system, computer device and readable storage medium

Publications (2)

Publication Number Publication Date
CN112711699A CN112711699A (en) 2021-04-27
CN112711699B true CN112711699B (en) 2023-04-07

Family

ID=75540551

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911016451.3A Active CN112711699B (en) 2019-10-24 2019-10-24 User division method, system, computer device and readable storage medium

Country Status (1)

Country Link
CN (1) CN112711699B (en)

Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104424302A (en) * 2013-09-04 2015-03-18 阿里巴巴集团控股有限公司 Method and device for matching homogeneous data objects
CN105631749A (en) * 2015-12-24 2016-06-01 成都陌云科技有限公司 User portrait calculation method based on statistical data
CN106055617A (en) * 2016-05-26 2016-10-26 乐视控股(北京)有限公司 Data pushing method and device
CN106202391A (en) * 2016-07-08 2016-12-07 深圳市中北明夷科技有限公司 The automatic classification method of a kind of user's community and device
CN106415640A (en) * 2014-06-27 2017-02-15 英特尔公司 Social and contextually appropriate recommendation system
CN107688984A (en) * 2017-07-27 2018-02-13 上海壹账通金融科技有限公司 Product information method for pushing, device, storage medium and computer equipment
CN108108419A (en) * 2017-12-15 2018-06-01 百度在线网络技术(北京)有限公司 A kind of information recommendation method, device, equipment and medium
CN108197224A (en) * 2017-12-28 2018-06-22 广州虎牙信息科技有限公司 User group sorting technique, storage medium and terminal
CN108269196A (en) * 2017-12-01 2018-07-10 优视科技有限公司 Add in the method, apparatus and computer equipment of network social association
CN108509551A (en) * 2018-03-19 2018-09-07 西北大学 A kind of micro blog network key user digging system under the environment based on Spark and method
CN108629054A (en) * 2018-06-26 2018-10-09 深圳市爱的网络科技有限公司 Community method for pushing, device, computer installation and computer readable storage medium
CN109033408A (en) * 2018-08-03 2018-12-18 腾讯科技(深圳)有限公司 Information-pushing method and device, computer readable storage medium, electronic equipment
CN109885674A (en) * 2019-02-14 2019-06-14 腾讯科技(深圳)有限公司 A kind of determination of theme label, information recommendation method and device

Patent Citations (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104424302A (en) * 2013-09-04 2015-03-18 阿里巴巴集团控股有限公司 Method and device for matching homogeneous data objects
CN106415640A (en) * 2014-06-27 2017-02-15 英特尔公司 Social and contextually appropriate recommendation system
CN105631749A (en) * 2015-12-24 2016-06-01 成都陌云科技有限公司 User portrait calculation method based on statistical data
CN106055617A (en) * 2016-05-26 2016-10-26 乐视控股(北京)有限公司 Data pushing method and device
CN106202391A (en) * 2016-07-08 2016-12-07 深圳市中北明夷科技有限公司 The automatic classification method of a kind of user's community and device
CN107688984A (en) * 2017-07-27 2018-02-13 上海壹账通金融科技有限公司 Product information method for pushing, device, storage medium and computer equipment
CN108269196A (en) * 2017-12-01 2018-07-10 优视科技有限公司 Add in the method, apparatus and computer equipment of network social association
CN108108419A (en) * 2017-12-15 2018-06-01 百度在线网络技术(北京)有限公司 A kind of information recommendation method, device, equipment and medium
CN108197224A (en) * 2017-12-28 2018-06-22 广州虎牙信息科技有限公司 User group sorting technique, storage medium and terminal
CN108509551A (en) * 2018-03-19 2018-09-07 西北大学 A kind of micro blog network key user digging system under the environment based on Spark and method
CN108629054A (en) * 2018-06-26 2018-10-09 深圳市爱的网络科技有限公司 Community method for pushing, device, computer installation and computer readable storage medium
CN109033408A (en) * 2018-08-03 2018-12-18 腾讯科技(深圳)有限公司 Information-pushing method and device, computer readable storage medium, electronic equipment
CN109885674A (en) * 2019-02-14 2019-06-14 腾讯科技(深圳)有限公司 A kind of determination of theme label, information recommendation method and device

Also Published As

Publication number Publication date
CN112711699A (en) 2021-04-27

Similar Documents

Publication Publication Date Title
CN110321422B (en) Method for training model on line, pushing method, device and equipment
CN108921221B (en) User feature generation method, device, equipment and storage medium
CN110704739B (en) Resource recommendation method and device and computer storage medium
WO2018103718A1 (en) Application recommendation method and apparatus, and server
CN104573054B (en) A kind of information-pushing method and equipment
CN110267097A (en) Video pushing method, device and electronic equipment based on characteristic of division
CN102426577A (en) Information processing apparatus, information processing system, information processing method, and program
CN110475132A (en) Direct broadcasting room kind identification method, device and data processing equipment
CN111159563A (en) Method, device and equipment for determining user interest point information and storage medium
WO2023024408A1 (en) Method for determining feature vector of user, and related device and medium
CN113963303A (en) Image processing method, video recognition method, device, equipment and storage medium
CN111737577A (en) Data query method, device, equipment and medium based on service platform
CN114780606B (en) Big data mining method and system
CN114330752A (en) Ranking model training method and ranking method
CN113947131A (en) Model training method, device, equipment and storage medium
CN112711699B (en) User division method, system, computer device and readable storage medium
JP7402492B2 (en) Comment evaluation system, comment evaluation method, and program
CN110781211B (en) Data analysis method and device
CN110472151B (en) Media data pushing method, device and computer readable storage medium
CN111683280A (en) Video processing method and device and electronic equipment
CN116955788A (en) Method, device, equipment, storage medium and program product for processing content
CN110569447A (en) network resource recommendation method and device and storage medium
CN113886690A (en) Movie recommendation method and device based on user behavior, electronic equipment and medium
CN113886585A (en) Item recommendation method, computer device and computer-readable storage medium
CN114491093A (en) Multimedia resource recommendation and object representation network generation method and device

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant