CN112395500B - Content data recommendation method, device, computer equipment and storage medium - Google Patents
Content data recommendation method, device, computer equipment and storage medium Download PDFInfo
- Publication number
- CN112395500B CN112395500B CN202011285730.2A CN202011285730A CN112395500B CN 112395500 B CN112395500 B CN 112395500B CN 202011285730 A CN202011285730 A CN 202011285730A CN 112395500 B CN112395500 B CN 112395500B
- Authority
- CN
- China
- Prior art keywords
- order
- crowd
- data
- model
- attribute data
- 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
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
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- 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)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to the field of data processing, and discloses a content data recommendation method, a device, computer equipment and a storage medium, wherein the method comprises the following steps: preprocessing the acquired user data to obtain data to be recommended; inputting consumption attribute data, social attribute data and access attribute data corresponding to a user into a content preference model, and inputting flow service attribute data into a scene recommendation model; extracting crowd characteristics through a first-order crowd clustering model to obtain a first-order crowd classification result, and performing scene adaptation through a scene recommendation model to obtain a theme scene; performing index analysis on the first-order crowd classification result and the access attribute data through a second-order index subdivision model to determine crowd preference tags; determining a content recommendation label according to the crowd preference label and the theme scene; content data is acquired and recommended to the user. The invention realizes crowd feature extraction, index analysis and scene adaptation on the user data and accurately recommends to the user.
Description
Technical Field
The present invention relates to the field of data processing of big data, and in particular, to a content data recommendation method, apparatus, computer device, and storage medium.
Background
With the rapid development of the mobile internet, people are increasingly popularizing to acquire the content information wanted by themselves from the mobile internet through the APP in the mobile terminal, but with the rapid development of the internet, the information quantity is also greatly increased, which can lead users to be unable to acquire the information really needed by themselves from the APP quickly when facing a large amount of information, so that the use rate of the APP is reduced. A better solution to this problem is to introduce a recommendation method that can recommend content actually of interest to the user in a large amount of information so that the user obtains content information actually preferred by the user from the recommended content.
Disclosure of Invention
The invention provides a content data recommendation method, a device, computer equipment and a storage medium, which realize crowd feature extraction, index analysis and scene adaptation of user data, determine content recommendation labels of users, automatically match content data, recommend the content data to the users, accurately recommend the content data to the users, improve experience satisfaction of the users and improve effectiveness of content data recommendation.
A content data recommendation method, comprising:
acquiring user data of a user, and preprocessing the user data to obtain data to be recommended; the data to be recommended comprises consumption attribute data, social attribute data, access attribute data and flow service attribute data;
inputting the consumption attribute data, the social attribute data and the access attribute data corresponding to a user into a content preference model, and inputting the flow service attribute data into a scene recommendation model; the content preference model is a multi-order model based on a two-step clustering method and a decision tree; the content preference model comprises a first-order crowd clustering model and a second-order index subdivision model;
extracting crowd characteristics of the consumption attribute data and the social attribute data through the first-order crowd clustering model to obtain a first-order crowd classification result corresponding to the user, and performing scene adaptation on the flow service attribute data through the scene recommendation model to obtain a theme scene corresponding to the user;
performing index analysis on the first-order crowd classification result and the access attribute data through the second-order index subdivision model, and determining crowd preference labels corresponding to the users;
Determining a content recommendation label corresponding to the user according to the crowd preference label and the theme scene corresponding to the user;
and acquiring content data matched with the content recommendation label from a content database, and recommending the acquired content data to the user.
A content data recommendation apparatus comprising:
the acquisition module is used for acquiring user data of a user, preprocessing the user data and obtaining data to be recommended; the data to be recommended comprises consumption attribute data, social attribute data, access attribute data and flow service attribute data;
the input module is used for inputting the consumption attribute data, the social attribute data and the access attribute data corresponding to the user into a content preference model and inputting the flow service attribute data into a scene recommendation model; the content preference model is a multi-order model based on a two-step clustering method and a decision tree; the content preference model comprises a first-order crowd clustering model and a second-order index subdivision model;
the identification module is used for extracting crowd characteristics of the consumption attribute data and the social attribute data through the first-order crowd clustering model to obtain a first-order crowd classification result corresponding to the user, and simultaneously carrying out scene adaptation on the flow service attribute data through the scene recommendation model to obtain a theme scene corresponding to the user;
The analysis module is used for carrying out index analysis on the first-order crowd classification result and the access attribute data through the second-order index subdivision model, and determining crowd preference labels corresponding to the users;
the determining module is used for determining a content recommendation label corresponding to the user according to the crowd preference label and the theme scene corresponding to the user;
and the recommending module is used for acquiring the content data matched with the content recommending label from the content database and recommending the acquired content data to the user.
A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, the processor implementing the steps of the content data recommendation method described above when the computer program is executed.
A computer readable storage medium storing a computer program which, when executed by a processor, implements the steps of the content data recommendation method described above.
According to the content data recommendation method, the device, the computer equipment and the storage medium, user data of a user are obtained, and the user data are preprocessed to obtain data to be recommended, wherein the data comprise consumption attribute data, social attribute data, access attribute data and flow service attribute data; inputting the consumption attribute data, the social attribute data and the access attribute data corresponding to a user into a content preference model, and inputting the flow service attribute data into a scene recommendation model; extracting crowd characteristics of the consumption attribute data and the social attribute data through the first-order crowd clustering model to obtain a first-order crowd classification result corresponding to the user, and performing scene adaptation on the flow service attribute data through the scene recommendation model to obtain a theme scene corresponding to the user; performing index analysis on the first-order crowd classification result and the access attribute data through the second-order index subdivision model, and determining crowd preference labels corresponding to the users; determining a content recommendation label corresponding to the user according to the crowd preference label and the theme scene; the content data matched with the content recommendation label is obtained from the content database, and the obtained content data is recommended to the user, so that the purposes of carrying out crowd feature extraction, index analysis and scene adaptation on the user data through a content preference model and a scene recommendation model, determining the content recommendation label of the user, automatically matching the content data, recommending the content data to the user, accurately recommending the content data to the user, improving the accuracy of content data recommendation, recommending the preferred content data to the user, avoiding the display of disliked content data to the user, improving the experience satisfaction degree of the user, and improving the effectiveness of content data recommendation are achieved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the description of the embodiments of the present invention will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a schematic view of an application environment of a content data recommendation method according to an embodiment of the present invention;
FIG. 2 is a flowchart of a content data recommendation method according to an embodiment of the present invention;
FIG. 3 is a flowchart of step S30 of a content data recommendation method according to an embodiment of the present invention;
FIG. 4 is a flowchart of step S303 of a content data recommendation method according to an embodiment of the present invention;
FIG. 5 is a flowchart of step S304 of a content data recommendation method according to an embodiment of the present invention;
FIG. 6 is a flowchart of step S40 of a content data recommendation method according to an embodiment of the present invention;
FIG. 7 is a flowchart of step S401 of a content data recommendation method according to an embodiment of the present invention;
FIG. 8 is a flowchart of step S403 of the content data recommendation method according to an embodiment of the present invention;
FIG. 9 is a schematic block diagram of a content data recommendation device in an embodiment of the present invention;
FIG. 10 is a schematic diagram of a computer device in accordance with an embodiment of the invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are some, but not all embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The content data recommendation method provided by the invention can be applied to an application environment as shown in fig. 1, wherein a client (computer equipment) communicates with a server through a network. Among them, clients (computer devices) include, but are not limited to, personal computers, notebook computers, smartphones, tablet computers, cameras, and portable wearable devices. The server may be implemented as a stand-alone server or as a server cluster composed of a plurality of servers.
In one embodiment, as shown in fig. 2, a content data recommendation method is provided, and the technical scheme mainly includes the following steps S10-S60:
S10, acquiring user data of a user, and preprocessing the user data to obtain data to be recommended; the data to be recommended comprises consumption attribute data, social attribute data, access attribute data and flow service attribute data.
Understandably, when a user needs to acquire content information wanted by the user in application software of a mobile terminal of the user, a recommendation command is triggered on an interface of the application software, the user data is acquired, the user data is data of a relevant attribute corresponding to the user in a server corresponding to the application software, the user data comprises data of attributes such as consumption attribute, social attribute, access attribute and flow service attribute corresponding to the user, preprocessing is a process of performing regular expression processing, missing value supplementation or depolarization value processing on the user data, the regular expression processing is used for uniformly converting data of one attribute into data required by a data format corresponding to the attribute, the missing value supplementation is used for uniformly converting data of a data existence space corresponding to the attribute into filling data corresponding to the attribute, the depolarization value processing is used for uniformly converting a data statistics of one attribute exceeding or being lower than a limit value set by the attribute into a limit value adjacent to the data statistics, the user data is determined to be used as the data to be used for waiting after the preprocessing, the data is used for processing, the data is related to the data of the social attribute, the data is data of the social attribute, the data is related to the data, the data is the data of the attribute, the data is related to the attribute, the data is the data of the service, the service is related to the attribute, the data is the data of the user has a service, traffic packages, and the like.
S20, inputting the consumption attribute data, the social attribute data and the access attribute data corresponding to the user into a content preference model, and inputting the flow service attribute data into a scene recommendation model; the content preference model is a multi-order model based on a two-step clustering method and a decision tree; the content preference model comprises a first-order crowd clustering model and a second-order index subdivision model.
Understandably, the content preference model is a multi-order model which is based on a two-step clustering method and a decision tree algorithm and is constructed; the content preference model comprises a first-order crowd clustering model and a second-order index subdivision model, the two-step clustering method is a method of carrying out preliminary clustering through a hierarchical clustering or density clustering method to obtain a preliminary clustering result, and then a segmentation clustering method is applied to carry out secondary clustering from the preliminary clustering result, the decision tree algorithm is an algorithm of establishing a decision model according to the data attribute by adopting a tree structure, the content preference model can automatically generate crowd preference labels of the user according to the consumption attribute data, social attribute data and access attribute data of the user, the crowd preference labels mark out the user preference, for example, the crowd preference labels comprise the wisdom, contemporary, youth literature, life, socioepity, fantasy and the like in the cartoon field, the love and the like, the fantasy, the knight and the true man in the video field and the like, the scene recommendation model is a trained neural network model, the network structure of the scene recommendation model can be set according to requirements, for example, the network structure of a BP neural network model, the LSTM neural network model and the like can automatically identify a scene with a theme which is suitable for a user with a large flow according to the required theme, for a scene, and the scene of the user has a large flow, and the scene of the required theme can be automatically identified according to the video service data.
S30, extracting crowd characteristics of the consumption attribute data and the social attribute data through the first-order crowd clustering model to obtain a first-order crowd classification result corresponding to the user, and performing scene adaptation on the flow service attribute data through the scene recommendation model to obtain a theme scene corresponding to the user.
Understandably, the crowd feature is a relevant feature of the classification of the extracted crowd, the crowd feature is extracted as a process of extracting a feature of an attribute difference between crowds, the crowd feature extraction may include crowd feature exploration, analysis in a decision tree algorithm and path restoration, the crowd feature exploration includes crowd density clustering and crowd feature clustering, the first-order crowd classification result is a crowd type, namely, a crowd type in a first-order crowd type in the whole, and the scene is adapted to be identified through a process of carrying out convolution on the traffic service attribute data, and the adaptation and the theme scene of the user are automatically identified through the traffic service attribute data.
The first-order crowd clustering model can be a clustering model based on a density clustering and decision tree algorithm, and can also be a clustering model based on hierarchical clustering and BP neural network, and can automatically extract crowd characteristics according to consumption attribute data and social attribute data, classify according to the extracted crowd characteristics and output crowd types of users.
In one embodiment, as shown in fig. 3, before the step S30, that is, before the crowd feature extraction is performed on the consumption attribute data and the social attribute data by the first-order crowd clustering model, the method includes:
s301, acquiring a sample data set.
Understandably, the sample data set includes consumption attribute sample data, social attribute sample data, and access attribute sample data.
S302, screening the sample data set according to the first-order attribute, and screening out the first-order attribute data set.
Understandably, the first-order attributes include a consumption attribute, which is an attribute related to user consumption, and a social attribute, which is a social basic identity attribute, a held terminal attribute, an enjoyment business attribute, and the like, of the user.
S303, inputting the first-order attribute data set into a two-step clustering model, and carrying out crowd characteristic exploration on the first-order attribute data through the two-step clustering model to obtain a first-order crowd clustering result.
Understandably, the two-step clustering model is a model based on a two-step clustering method, the two-step clustering method is a method of performing preliminary clustering through hierarchical clustering or density clustering to obtain a preliminary clustering result, and then a method of performing secondary clustering by using a segmentation clustering method from the preliminary clustering result, wherein the crowd feature exploration is a process of performing standardization processing, crowd density clustering and crowd feature clustering on the first-order attribute dataset, and the characteristics of attribute similarity and dissimilarity among crowds are explored, so that the first-order crowd clustering result is obtained, and the first-order crowd clustering result is a crowd type which is primarily explored, such as a 9-class crowd type.
In an embodiment, as shown in fig. 4, in step S303, the performing, by using the two-step clustering model, crowd feature exploration on the first-order attribute data to obtain a first-order crowd clustering result includes:
s3031, carrying out standardized processing on the first-order attribute data set through the two-step clustering model to obtain first-order attribute data to be processed; the two-step clustering model comprises a density clustering model and a K-means clustering model.
It is to be appreciated that the normalization process is a process of performing the regular expression process, the missing value supplementing process, the depolarization process, the one-hot transcoding process, and the regularization process on the first-order attribute data set, where the one-hot transcoding process is also called one-bit efficient coding, N-bit state registers are mainly used to encode N states, each state is assigned an integer value, the regularization process is a process of taking a sum of absolute values of vectors of each sample as a norm, removing the norm with each vector, and obtaining a vector regularized by the sample, or a process of squaring and then dividing the vector of each sample as a norm and then dividing the vector, so that the normalization process is performed on the first-order attribute data set to obtain the first-order attribute data to be processed.
The two-step clustering model comprises a density clustering model and a K-means clustering model, and the first-order attribute data to be processed are data provided for clustering of the two-step clustering model.
S3032, using a DBSCAN algorithm, and carrying out crowd density clustering on the first-order attribute data to be processed through the density clustering model to obtain a transitional clustering data result.
Understandably, the DBSCAN (Density-basedClustering Method, density-based clustering algorithm) algorithm is an algorithm for determining each type of region by the Density condition of each region, isolating abnormal values, and determining the abnormal values as a class, the crowd Density clustering is a clustering process for determining all crowd types by using the DBSCAN algorithm, and the transitional clustering data result is crowd types obtained after the crowd Density clustering, such as 8 types of crowd types, wherein all abnormal values are classified as one type of crowd type (abnormal type).
The density clustering model is a model for clustering and distinguishing crowd types by using the DBSCAN algorithm.
S3033, using a K-means algorithm, and carrying out crowd feature clustering on the transition clustering data result through the K-means clustering model to obtain the first-order crowd clustering result.
The K-means algorithm is a partition clustering algorithm taking an average value as a center of a class, the partition clustering algorithm randomly selects objects from a data set as prototypes of clusters, and distributes other objects to the most similar (namely, the most recent class) represented by the prototypes, the K-means clustering model is a model for determining crowd types by using the K-means algorithm to cluster the transition cluster data results, and the crowd feature clustering is a clustering process for determining all crowd types based on the transition cluster data results by using the K-means algorithm, wherein the first-order crowd clustering result comprises crowd types corresponding to abnormal types in the transition cluster data results.
The method realizes the standardized processing of the first-order attribute data set through the two-step clustering model to obtain first-order attribute data to be processed; using a DBSCAN algorithm, and carrying out crowd density clustering on the first-order attribute data to be processed through the density clustering model to obtain a transitional clustering data result; and carrying out crowd feature clustering on the transition clustering data result through the K-means clustering model by using a K-means algorithm to obtain the first-order crowd clustering result, so that crowd density clustering and crowd feature clustering are carried out through preprocessing, a DBSCAN algorithm and the K-means algorithm, thereby obtaining the first-order crowd clustering result and improving the accuracy of crowd classification.
S304, analyzing and path restoring the first-order crowd clustering result and the first-order effective data set through a decision tree algorithm, and refining at least one classification variable corresponding to the first-order crowd clustering result.
The decision tree algorithm is an algorithm for establishing a decision model by adopting a tree structure according to the data attribute, the analysis and the path restore to a back-pushing process in the decision tree algorithm, the decision nodes of each attribute of the sample user are analyzed through analyzing the corresponding relation between the first-order effective data set and the first-order crowd clustering result, and the paths of the passing decision nodes are pushed back, so that the paths from the first-order effective data set decision to the crowd types in the first-order crowd clustering result are restored, and the variables corresponding to the decision nodes of the attribute with the threshold value of the passing times of the path nodes are extracted and determined to be the classification variables.
In an embodiment, as shown in fig. 5, in step S304, that is, the step of analyzing and path-restoring the first-order crowd clustering result and the first-order valid data set by using a decision tree algorithm, extracting at least one classification variable corresponding to the first-order crowd clustering result includes:
S3041, associating first-order crowd types corresponding to the same sample users with first-order effective data, and determining the associated first-order effective data set as a decision data set; the first-order crowd clustering result comprises the first-order crowd type corresponding to the sample user in the first-order effective data set; the first-order valid data set includes the first-order valid data in one-to-one correspondence with the sample users.
Understandably, the crowd types of the sample users in the first-order effective data set can be partitioned through the two-step clustering model, so that the crowd type of the first-order effective data corresponding to the sample users one by one, namely, the first-order crowd type, can be determined, the correlated first-order effective data set is determined as a decision data set, and all the first-order effective data are correlated.
S3042, inputting the decision data set into a decision back-pushing model containing initial variable parameters;
the decision back-push model is a model for carrying out back-push on the decision data according to the tree structure of the decision tree to identify the variable parameters of crowd types.
S3043, analyzing the decision data set through the decision back-pushing model by using a decision tree algorithm, and updating the initial variable parameters.
The decision tree algorithm is an algorithm for establishing a decision model according to the attribute of data by adopting a tree structure, namely dividing the decision data set according to the data characteristics of the decision data set until all the characteristics are divided or the crowd types of all the data of the divided data subset are the same, and then gathering according to the first-order crowd type associated with the first-order effective data in the decision data set, continuously deducing and analyzing variable parameters capable of dividing the first-order crowd type, updating initial variable parameters until the initial variable parameters are completely gathered, and determining the initial variable parameters at the moment as the updated initial variable parameters.
S3044, carrying out path reduction according to the updated initial variable parameters, and refining the classification variable corresponding to the first-order crowd clustering result.
It is to be understood that the path is restored to a divided path for restoring each first-order valid data, whether the first-order crowd type corresponding to the path can be reached is confirmed, after all paths are restored, the classification variable is determined according to the number of nodes where the paths overlap, the determination mode can be set according to requirements, for example, the number of each divided node is identified, the variable parameter in the node greater than or equal to the preset number is determined as the classification variable, or the variable parameter in the node greater than or equal to the average value of the number of all nodes passed is determined as the classification variable, and so on.
The invention realizes that the first-order effective data set after being correlated is determined as a decision data set by correlating the first-order crowd type corresponding to the same sample user with the first-order effective data; inputting the decision data set into a decision back-pushing model containing initial variable parameters; analyzing the decision data set through the decision back-pushing model by using a decision tree algorithm, and updating the initial variable parameters; and carrying out path restoration according to the updated initial variable parameters, and extracting the classification variables corresponding to the first-order crowd clustering result, so that the rule of subdivision can be analyzed and path restored through a decision back-push model, the classification variables are extracted, the classification variables can be extracted more scientifically by using a decision tree algorithm, and the quality and accuracy of crowd subdivision are improved.
S305, carrying out model reconstruction according to all the classification variables, the first-order clustering result and the first-order effective data set, constructing a first-order crowd clustering model, determining first-order crowd types corresponding to the first-order crowd clustering model, marking each first-order effective data in the first-order effective data set with the crowd type corresponding to the first-order effective data set, and obtaining a first-order data set; the first order demographic categories include at least one of the demographic types.
Understandably, the model reconstruction is performed on all the classification variables, the first-order clustering result and the first-order valid data set again through a decision tree algorithm, so as to construct a first-order crowd clustering model, wherein the first-order crowd type comprises at least one crowd type, such as 11 crowd types, and each first-order valid data in the first-order valid data set is marked with the corresponding crowd type.
The invention realizes the aim of acquiring a sample data set; screening the sample data set according to the first-order attribute to obtain a first-order attribute data set; inputting the first-order attribute data set into a two-step clustering model, and performing crowd characteristic exploration on the first-order attribute data through the two-step clustering model to obtain a first-order crowd clustering result; analyzing the first-order crowd clustering result and the first-order effective data set through a decision tree algorithm, and carrying out path restoration to extract at least one classification variable corresponding to the first-order crowd clustering result; and carrying out model reconstruction according to all the classification variables, the first-order clustering result and the first-order effective data set to construct a first-order crowd clustering model, and determining the first-order crowd types corresponding to the first-order crowd clustering model, so that crowd characteristic exploration and path restoration are carried out through a two-step clustering model and a decision tree algorithm, classification variables are analyzed, a first-order crowd clustering model is constructed, the first-order crowd clustering model can be accurately constructed, and the crowd classification accuracy is improved.
And S40, performing index analysis on the first-order crowd classification result and the access attribute data through the second-order index subdivision model, and determining crowd preference labels corresponding to the users.
Understandably, the first-order crowd classification result is a crowd type, namely one crowd type in the first-order crowd types in the whole text, the second-order index subdivision model is a clustering model of the built user subdivision, the second-order index subdivision model can analyze indexes according to the crowd type in the first-order crowd classification result and the access attribute data, and can analyze and match crowd preference labels of users, and the crowd preference labels label marks preferences of the users.
In one embodiment, as shown in fig. 6, before the step S40, that is, before the performing, by the second-order index subdivision model, index analysis on the first-order crowd classification result and the access attribute data, the method includes:
s401, merging to generate a second-order attribute data set according to the first-order data set and access attribute sample data in the sample data set.
Understandably, the access attribute sample data in the sample data set is correspondingly added to the first-order data set, that is, the access attribute sample data of one user is added to the first-order valid data in the first-order data set corresponding to the user, and the access attribute sample data may be inserted after the first-order valid data, so as to combine to obtain the second-order attribute data set.
In one embodiment, as shown in fig. 7, in the step S401, that is, the step of merging the first-order data set and the access attribute sample data in the sample data set to generate a second-order attribute data set includes:
s4011, randomly extracting fields from the access attribute sample data to extract attribute data to be processed.
Understandably, the random extraction field is a field in all the access attribute sample data extracted randomly, so that analysis can be performed through scattered data distribution conditions, analysis can be performed on access behaviors of users more objectively, and the extracted field data is determined to be the attribute data to be processed.
S4012, carrying out missing value processing and extremum processing on the attribute data to be processed to obtain the attribute data to be added.
It is understood that the missing value processing and the extremum processing are performed on all the attribute data to be processed, the missing value processing includes deleting the data containing the missing value and the possible value interpolation missing value, namely deleting the data containing the missing value for the attribute data of a part of fields, supplementing the data in a mode of performing the possible value interpolation missing value for the attribute data of a part of fields, and the extremum processing is a processing procedure of removing or replacing the data of one attribute exceeding or being lower than the limit value set by the attribute into unified data.
S4013, correspondingly adding the attribute data to be added to the first-order data set, and generating the second-order attribute data set.
Understandably, the attribute data to be added is inserted after the first-order valid data, so that the second-order attribute data sets are obtained by merging.
And S402, extracting index features of the second-order attribute data set through a preference behavior model to obtain at least one comprehensive index variable.
The index feature extraction is a process of calculating the contribution degree of each index according to the second-order attribute data set, and extracting the index reaching the threshold requirement, wherein the contribution degree is the contribution degree, namely the occupation degree, of the index in the access behavior data of the user.
S403, carrying out sectional analysis on the second-order attribute data set according to the first-order crowd classification result and all the comprehensive index variables, and constructing a second-order index subdivision model.
Understandably, according to the first-order crowd classification result and all the comprehensive index variables, each first-order crowd type is subdivided into sections corresponding to the comprehensive index variables one by one according to the comprehensive index variables, analysis of each section is performed through the sections in which the second-order attribute data set falls, and the segmentation analysis is to perform a duty ratio and weight analysis process on each divided section, that is, to perform adjacent section merging or splitting on each section, so that the duty ratio of each processed section is greater than or equal to a preset duty ratio and the weight is greater than a preset weight, and the second-order index subdivision model is constructed according to the sections reaching the requirements.
The invention realizes that a second-order attribute data set is generated by combining according to the first-order data set and access attribute sample data in the sample data set; extracting index features of the second-order attribute data set through a preference behavior model to obtain at least one comprehensive index variable; and according to the first-order crowd classification result and all the comprehensive index variables, carrying out sectional analysis on the second-order attribute data set and constructing a second-order index subdivision model, so that index feature extraction through a preference behavior model is realized, and after sectional analysis, the second-order index subdivision model is constructed, thereby accurately, scientifically and objectively constructing the second-order index subdivision model and improving the accuracy and reliability of crowd subdivision.
In an embodiment, as shown in fig. 8, in step S403, the step of performing a segmentation analysis on the second-order attribute dataset according to the first-order crowd classification result and all the comprehensive index variables, and constructing a second-order index subdivision model includes:
s4031, performing feature analysis and dimension reduction on all the comprehensive index variables to obtain main component index variables.
Understandably, the comprehensive index variable includes a plurality of indexes, the number of the indexes is too large, the user feature extraction and the user subdivision cannot be directly researched, the indexes are required to be integrated, correlation of attribute data is further analyzed, the indexes are subjected to dimension reduction processing, the feature analysis is to calculate similarity of the indexes in each dimension, the similarity value of each index in each dimension is analyzed, namely, the maximum similarity value of each index and each dimension is analyzed, and the two-dimension merging and classifying process is performed by combining the contribution degree of each index, so that the total contribution degree and the average contribution degree tolerance of the two-dimension merged dimension are minimum, the dimension merged by the two dimensions is determined to be a progressive dimension, and the total contribution degree is the sum of the contribution degrees of the indexes in the progressive dimension.
The dimension reduction process is a process of setting a weight parameter for each index in a step dimension, wherein the weight parameter is the duty ratio of the index in the step dimension corresponding to the index, namely the duty ratio of the contribution degree of the index in the total contribution degree of the step dimension, calculating the weight average value of the weight parameters in all the step dimensions, merging the index corresponding to the weight parameter larger than the weight average value with the step dimension corresponding to the index, and determining the combined weight average value as the main component index variable, so that the dimension reduction can be carried out on a plurality of step dimensions into a plurality of representative main component index variables, and the main component index variables indicate main factors of user content preference.
S4032, associating the crowd types corresponding to the same user with second-order attribute data, and determining the second-order attribute data set after association as a data set to be subdivided; the first-order crowd classification result comprises the crowd type corresponding to the user; the second order attribute data set includes the second order attribute data in one-to-one correspondence with the user.
Understandably, associating the crowd type corresponding to the same user with the second-order attribute data is equivalent to assigning a label of the crowd type to the second-order attribute data, and the first-order crowd classification result includes the crowd type corresponding to the user.
S4033, carrying out sectional analysis on the data set to be subdivided according to the main component index variable, and constructing the second-order index subdivision model.
The segment analysis is an analysis process of dividing each second-order attribute data in the data set to be subdivided into segments of each principal component index variable, namely a learning process according to the clustering degree between the second-order attribute data and the principal component index variable, wherein the learning mode is unsupervised clustering learning, so that the second-order index subdivision model is constructed.
The invention realizes that the main component index variable is obtained by carrying out feature analysis and dimension reduction treatment on all the comprehensive index variables; correlating the crowd types and the second-order attribute data corresponding to the same user, and determining the correlated second-order attribute data set as a data set to be subdivided; according to the main component index variable, the data set to be subdivided is subjected to sectional analysis, and the second-order index subdivision model is constructed, so that the crowd can be subdivided more directly through feature analysis and dimension reduction processing, and the crowd subdivision reliability and accuracy are improved.
S50, determining content recommendation labels corresponding to the users according to the crowd preference labels and the theme scenes corresponding to the users.
Understandably, according to the determined crowd preference label and the determined theme scene corresponding to the user, mapping the content recommendation label matched with both the crowd preference label and the theme scene, where the content recommendation label is a label to which content data recommended to the user belongs, and the content recommendation label may be set according to requirements, for example, the content label may be a video fantasy, a video martial arts, or the like.
S60, acquiring content data matched with the content recommendation label from a content database, and recommending the acquired content data to the user.
Understandably, the content database is all content data stored in a period of time or on the same day stored in the server corresponding to the application software, the content data is content information on the mobile internet, the content data is marked with a content tag, the content data corresponding to the content tag matched with the content recommendation tag is found from the content database, the found content data is obtained, and is recommended to an interface of the application software of the mobile terminal corresponding to the user for the user to view.
The invention realizes preprocessing the user data by acquiring the user data of the user to obtain the data to be recommended comprising consumption attribute data, social attribute data, access attribute data and flow service attribute data; inputting the consumption attribute data, the social attribute data and the access attribute data corresponding to a user into a content preference model, and inputting the flow service attribute data into a scene recommendation model; extracting crowd characteristics of the consumption attribute data and the social attribute data through the first-order crowd clustering model to obtain a first-order crowd classification result corresponding to the user, and performing scene adaptation on the flow service attribute data through the scene recommendation model to obtain a theme scene corresponding to the user; performing index analysis on the first-order crowd classification result and the access attribute data through the second-order index subdivision model, and determining crowd preference labels corresponding to the users; determining a content recommendation label corresponding to the user according to the crowd preference label and the theme scene; the content data matched with the content recommendation label is obtained from the content database, and the obtained content data is recommended to the user, so that the purposes of carrying out crowd feature extraction, index analysis and scene adaptation on the user data through a content preference model and a scene recommendation model, determining the content recommendation label of the user, automatically matching the content data, recommending the content data to the user, accurately recommending the content data to the user, improving the accuracy of content data recommendation, recommending the preferred content data to the user, avoiding the display of disliked content data to the user, improving the experience satisfaction degree of the user, and improving the effectiveness of content data recommendation are achieved.
In an embodiment, a content data recommendation device is provided, where the content data recommendation device corresponds to the content data recommendation method in the above embodiment one by one. As shown in fig. 9, the content data recommendation apparatus includes an acquisition module 11, an input module 12, an identification module 13, an analysis module 14, a determination module 15, and a recommendation module 16. The functional modules are described in detail as follows:
the acquisition module 11 is used for acquiring user data of a user, preprocessing the user data and obtaining data to be recommended; the data to be recommended comprises consumption attribute data, social attribute data, access attribute data and flow service attribute data;
an input module 12 for inputting the consumption attribute data, the social attribute data, and the access attribute data corresponding to a user into a content preference model, and simultaneously inputting the flow service attribute data into a scene recommendation model; the content preference model is a multi-order model based on a two-step clustering method and a decision tree; the content preference model comprises a first-order crowd clustering model and a second-order index subdivision model;
the recognition module 13 is configured to perform crowd feature extraction on the consumption attribute data and the social attribute data through the first-order crowd clustering model to obtain a first-order crowd classification result corresponding to the user, and simultaneously perform scene adaptation on the flow service attribute data through the scene recommendation model to obtain a theme scene corresponding to the user;
The analysis module 14 is configured to perform index analysis on the first-order crowd classification result and the access attribute data through the second-order index subdivision model, and determine crowd preference tags corresponding to the users;
a determining module 15, configured to determine a content recommendation label corresponding to the user according to the crowd preference label and the theme scene corresponding to the user;
and a recommending module 16, configured to acquire content data matched with the content recommendation tag from a content database, and recommend the acquired content data to the user.
For specific limitations of the content data recommendation device, reference may be made to the above limitation of the content data recommendation method, and no further description is given here. The respective modules in the content data recommendation apparatus described above may be implemented in whole or in part by software, hardware, and a combination thereof. The above modules may be embedded in hardware or may be independent of a processor in the computer device, or may be stored in software in a memory in the computer device, so that the processor may call and execute operations corresponding to the above modules.
In one embodiment, a computer device is provided, which may be a server, and the internal structure of which may be as shown in fig. 10. The computer device includes a processor, a memory, a network interface, and a database connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs, and a database. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage media. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a content data recommendation method.
In one embodiment, a computer device is provided that includes a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor implementing the content data recommendation method of the above embodiments when the computer program is executed by the processor.
In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored, which when executed by a processor implements the content data recommendation method in the above embodiments.
Those skilled in the art will appreciate that implementing all or part of the above described methods may be accomplished by way of a computer program stored on a non-transitory computer readable storage medium, which when executed, may comprise the steps of 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.
It will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of the functional units and modules is illustrated, and in practical application, the above-described functional distribution may be performed by different functional units and modules according to needs, i.e. the internal structure of the apparatus is divided into different functional units or modules to perform all or part of the above-described functions.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention, and are intended to be included in the scope of the present invention.
Claims (8)
1. A content data recommendation method, comprising:
acquiring user data of a user, and preprocessing the user data to obtain data to be recommended; the data to be recommended comprises consumption attribute data, social attribute data, access attribute data and flow service attribute data;
Inputting the consumption attribute data, the social attribute data and the access attribute data corresponding to a user into a content preference model, and inputting the flow service attribute data into a scene recommendation model; the content preference model is a multi-order model based on a two-step clustering method and a decision tree; the content preference model comprises a first-order crowd clustering model and a second-order index subdivision model;
extracting crowd characteristics of the consumption attribute data and the social attribute data through the first-order crowd clustering model to obtain a first-order crowd classification result corresponding to the user, and performing scene adaptation on the flow service attribute data through the scene recommendation model to obtain a theme scene corresponding to the user;
performing index analysis on the first-order crowd classification result and the access attribute data through the second-order index subdivision model, and determining crowd preference labels corresponding to the users;
determining a content recommendation label corresponding to the user according to the crowd preference label and the theme scene corresponding to the user;
acquiring content data matched with the content recommendation label from a content database, and recommending the acquired content data to the user;
Before the crowd feature extraction is performed on the consumption attribute data and the social attribute data through the first-order crowd clustering model, the method comprises the following steps:
acquiring a sample data set;
screening the sample data set according to the first-order attribute to obtain a first-order attribute data set;
inputting the first-order attribute data set into a two-step clustering model, and performing crowd characteristic exploration on the first-order attribute data through the two-step clustering model to obtain a first-order crowd clustering result;
analyzing the first-order crowd clustering result and the first-order effective data set through a decision tree algorithm, and carrying out path restoration to extract at least one classification variable corresponding to the first-order crowd clustering result;
performing model reconstruction according to all the classification variables, the first-order crowd clustering results and the first-order effective data set, constructing a first-order crowd clustering model, determining first-order crowd types corresponding to the first-order crowd clustering model, marking each first-order effective data in the first-order effective data set with the crowd type corresponding to the first-order effective data, and obtaining a first-order data set; the first order crowd category includes at least one of the crowd types;
Before the index analysis is performed on the first-order crowd classification result and the access attribute data through the second-order index subdivision model, the method comprises the following steps:
combining to generate a second-order attribute data set according to the first-order data set and the access attribute sample data in the sample data set;
extracting index features of the second-order attribute data set through a preference behavior model to obtain at least one comprehensive index variable;
and carrying out sectional analysis on the second-order attribute data set according to the first-order crowd classification result and all the comprehensive index variables, and constructing a second-order index subdivision model.
2. The content data recommendation method according to claim 1, wherein the performing crowd feature exploration on the first-order attribute data through the two-step clustering model to obtain a first-order crowd clustering result comprises:
carrying out standardized processing on the first-order attribute data set through the two-step clustering model to obtain first-order attribute data to be processed; the two-step clustering model comprises a density clustering model and a K-means clustering model;
using a DBSCAN algorithm, and carrying out crowd density clustering on the first-order attribute data to be processed through the density clustering model to obtain a transitional clustering data result;
And carrying out crowd characteristic clustering on the transition clustering data result through the K-means clustering model by using a K-means algorithm to obtain the first-order crowd clustering result.
3. The content data recommendation method according to claim 1, wherein analyzing and path-restoring the first-order crowd-clustered results and the first-order valid data set by a decision tree algorithm to extract at least one classification variable corresponding to the first-order crowd-clustered results comprises:
correlating the first-order crowd types corresponding to the same sample user with first-order effective data, and determining the correlated first-order effective data set as a decision data set; the first-order crowd clustering result comprises the first-order crowd type corresponding to the sample user in the first-order effective data set; the first-order valid data set comprises the first-order valid data which corresponds to the sample users one by one;
inputting the decision data set into a decision back-pushing model containing initial variable parameters;
analyzing the decision data set through the decision back-pushing model by using a decision tree algorithm, and updating the initial variable parameters;
and carrying out path reduction according to the updated initial variable parameters, and refining the classification variable corresponding to the first-order crowd clustering result.
4. The content data recommendation method according to claim 1, wherein said merging generates a second order attribute data set from access attribute sample data in said first order data set and said sample data set, comprising:
randomly extracting a field from the access attribute sample data to extract attribute data to be processed;
carrying out missing value processing and extremum processing on the attribute data to be processed to obtain attribute data to be added;
and correspondingly adding the attribute data to be added to the first-order data set to generate the second-order attribute data set.
5. The content data recommendation method according to claim 1, wherein said performing a segmentation analysis on said second-order attribute data set based on said first-order population classification result and all said comprehensive index variables, and constructing a second-order index subdivision model, comprises:
performing feature analysis and dimension reduction treatment on all the comprehensive index variables to obtain main component index variables;
correlating the crowd types and the second-order attribute data corresponding to the same user, and determining the correlated second-order attribute data set as a data set to be subdivided; the first-order crowd classification result comprises the crowd type corresponding to the user; the second-order attribute data set comprises the second-order attribute data in one-to-one correspondence with the users;
And carrying out sectional analysis on the data set to be subdivided according to the main component index variable, and constructing the second-order index subdivision model.
6. A content data recommendation apparatus, comprising:
the acquisition module is used for acquiring user data of a user, preprocessing the user data and obtaining data to be recommended; the data to be recommended comprises consumption attribute data, social attribute data, access attribute data and flow service attribute data;
the input module is used for inputting the consumption attribute data, the social attribute data and the access attribute data corresponding to the user into a content preference model and inputting the flow service attribute data into a scene recommendation model; the content preference model is a multi-order model based on a two-step clustering method and a decision tree; the content preference model comprises a first-order crowd clustering model and a second-order index subdivision model;
the identification module is used for extracting crowd characteristics of the consumption attribute data and the social attribute data through the first-order crowd clustering model to obtain a first-order crowd classification result corresponding to the user, and simultaneously carrying out scene adaptation on the flow service attribute data through the scene recommendation model to obtain a theme scene corresponding to the user;
The analysis module is used for carrying out index analysis on the first-order crowd classification result and the access attribute data through the second-order index subdivision model, and determining crowd preference labels corresponding to the users;
the determining module is used for determining a content recommendation label corresponding to the user according to the crowd preference label and the theme scene corresponding to the user;
the recommending module is used for acquiring content data matched with the content recommending label from a content database and recommending the acquired content data to the user;
the identification module is also used for:
acquiring a sample data set;
screening the sample data set according to the first-order attribute to obtain a first-order attribute data set;
inputting the first-order attribute data set into a two-step clustering model, and performing crowd characteristic exploration on the first-order attribute data through the two-step clustering model to obtain a first-order crowd clustering result;
analyzing the first-order crowd clustering result and the first-order effective data set through a decision tree algorithm, and carrying out path restoration to extract at least one classification variable corresponding to the first-order crowd clustering result;
Performing model reconstruction according to all the classification variables, the first-order crowd clustering results and the first-order effective data set, constructing a first-order crowd clustering model, determining first-order crowd types corresponding to the first-order crowd clustering model, marking each first-order effective data in the first-order effective data set with the crowd type corresponding to the first-order effective data, and obtaining a first-order data set; the first order crowd category includes at least one of the crowd types;
the analysis module is also configured to:
combining to generate a second-order attribute data set according to the first-order data set and the access attribute sample data in the sample data set;
extracting index features of the second-order attribute data set through a preference behavior model to obtain at least one comprehensive index variable;
and carrying out sectional analysis on the second-order attribute data set according to the first-order crowd classification result and all the comprehensive index variables, and constructing a second-order index subdivision model.
7. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the content data recommendation method according to any one of claims 1 to 5 when executing the computer program.
8. A computer readable storage medium storing a computer program, wherein the computer program when executed by a processor implements the content data recommendation method according to any one of claims 1 to 5.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011285730.2A CN112395500B (en) | 2020-11-17 | 2020-11-17 | Content data recommendation method, device, computer equipment and storage medium |
| PCT/CN2021/091067 WO2022105129A1 (en) | 2020-11-17 | 2021-04-29 | Content data recommendation method and apparatus, and computer device, and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011285730.2A CN112395500B (en) | 2020-11-17 | 2020-11-17 | Content data recommendation method, device, computer equipment and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112395500A CN112395500A (en) | 2021-02-23 |
| CN112395500B true CN112395500B (en) | 2023-09-05 |
Family
ID=74599815
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011285730.2A Active CN112395500B (en) | 2020-11-17 | 2020-11-17 | Content data recommendation method, device, computer equipment and storage medium |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN112395500B (en) |
| WO (1) | WO2022105129A1 (en) |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112395500B (en) * | 2020-11-17 | 2023-09-05 | 平安科技（深圳）有限公司 | Content data recommendation method, device, computer equipment and storage medium |
| CN113094578B (en) * | 2021-03-16 | 2023-03-31 | 平安普惠企业管理有限公司 | Deep learning-based content recommendation method, device, equipment and storage medium |
| CN113095570B (en) * | 2021-04-14 | 2023-06-06 | 上海市城市建设设计研究总院（集团）有限公司 | Bicycle riding path recommending method based on demand difference |
| CN113837859B (en) * | 2021-08-25 | 2024-05-14 | 天元大数据信用管理有限公司 | Image construction method for small and micro enterprises |
| CN116450925B (en) * | 2022-12-27 | 2023-12-15 | 深圳市网新新思软件有限公司 | User relationship analysis method and system based on artificial intelligence |
| CN116595342B (en) * | 2023-07-07 | 2023-09-29 | 北京数巅科技有限公司 | Crowd circling method, device and equipment and storage medium |
| CN117056613B (en) * | 2023-10-12 | 2024-06-04 | 中质国优测评技术(北京)有限公司 | Evaluation optimization method and system based on user joint preference |
| CN117763228A (en) * | 2023-12-11 | 2024-03-26 | 广州小白信息技术有限公司 | Creative expression dynamic adaptation method based on multi-culture framework |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109376759A (en) * | 2018-09-10 | 2019-02-22 | 平安科技（深圳）有限公司 | User information classification method, device, computer equipment and storage medium |
| WO2020114108A1 (en) * | 2018-12-04 | 2020-06-11 | 阿里巴巴集团控股有限公司 | Clustering result interpretation method and device |
Family Cites Families (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7165105B2 (en) * | 2001-07-16 | 2007-01-16 | Netgenesis Corporation | System and method for logical view analysis and visualization of user behavior in a distributed computer network |
| CN107798557B (en) * | 2017-09-30 | 2020-09-25 | 平安科技（深圳）有限公司 | Electronic device, service place recommendation method based on LBS data and storage medium |
| US20200073953A1 (en) * | 2018-08-30 | 2020-03-05 | Salesforce.Com, Inc. | Ranking Entity Based Search Results Using User Clusters |
| US10922725B2 (en) * | 2019-01-31 | 2021-02-16 | Salesforce.Com, Inc. | Automatic rule generation for recommendation engine using hybrid machine learning |
| CN111582932A (en) * | 2020-03-25 | 2020-08-25 | 平安壹钱包电子商务有限公司 | Inter-scene information pushing method and device, computer equipment and storage medium |
| CN111897861A (en) * | 2020-06-30 | 2020-11-06 | 苏宁金融科技(南京)有限公司 | Content recommendation method and device, computer equipment and storage medium |
| CN112395500B (en) * | 2020-11-17 | 2023-09-05 | 平安科技（深圳）有限公司 | Content data recommendation method, device, computer equipment and storage medium |
-
2020
- 2020-11-17 CN CN202011285730.2A patent/CN112395500B/en active Active
-
2021
- 2021-04-29 WO PCT/CN2021/091067 patent/WO2022105129A1/en active Application Filing
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109376759A (en) * | 2018-09-10 | 2019-02-22 | 平安科技（深圳）有限公司 | User information classification method, device, computer equipment and storage medium |
| WO2020114108A1 (en) * | 2018-12-04 | 2020-06-11 | 阿里巴巴集团控股有限公司 | Clustering result interpretation method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112395500A (en) | 2021-02-23 |
| WO2022105129A1 (en) | 2022-05-27 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112395500B (en) | Content data recommendation method, device, computer equipment and storage medium | |
| CN108921221B (en) | User feature generation method, device, equipment and storage medium | |
| CN108874992B (en) | Public opinion analysis method, system, computer equipment and storage medium | |
| CN107066616B (en) | Account processing method and device and electronic equipment | |
| WO2021027595A1 (en) | User portrait generation method and apparatus, computer device, and computer-readable storage medium | |
| CN109871490B (en) | Media resource matching method and device, storage medium and computer equipment | |
| CN109325118B (en) | Unbalanced sample data preprocessing method and device and computer equipment | |
| CN113204660B (en) | Multimedia data processing method, tag identification device and electronic equipment | |
| CN112035611B (en) | Target user recommendation method, device, computer equipment and storage medium | |
| CN110888911A (en) | Sample data processing method and device, computer equipment and storage medium | |
| CN112784168B (en) | Information push model training method and device, information push method and device | |
| CN110162664B (en) | Video recommendation method and device, computer equipment and storage medium | |
| CN110880006A (en) | User classification method and device, computer equipment and storage medium | |
| CN114491084B (en) | Self-encoder-based relation network information mining method, device and equipment | |
| CN111651666A (en) | User theme recommendation method and device, computer equipment and storage medium | |
| CN113704509A (en) | Multimedia recommendation method and device, electronic equipment and storage medium | |
| CN111291795A (en) | Crowd characteristic analysis method and device, storage medium and computer equipment | |
| CN114491093B (en) | Multimedia resource recommendation and object representation network generation method and device | |
| CN115222112A (en) | Behavior prediction method, behavior prediction model generation method and electronic equipment | |
| CN114741540A (en) | Multimedia sequence recommendation method, operation prediction model training method, device, equipment and storage medium | |
| CN115756821A (en) | Online task processing model training and task processing method and device | |
| CN113868516A (en) | Object recommendation method and device, electronic equipment and storage medium | |
| CN114021739B (en) | Business processing method, business processing model training device and electronic equipment | |
| CN114048392B (en) | Multimedia resource pushing method and device, electronic equipment and storage medium | |
| CN112001566B (en) | Optimization method, device, equipment and medium of fitness training model |
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 |