CN-113821596-B - Information recommendation method, device, computer equipment and storage medium

CN113821596BCN 113821596 BCN113821596 BCN 113821596BCN-113821596-B

Abstract

The invention discloses an information recommendation method, device, computer equipment and medium, which comprise the steps of acquiring registration data of a target object, generating an initialization tag of the target object based on the registration data, acquiring behavior data of the target object, determining interest words of the target object according to the behavior data, carrying out clustering processing based on the initialization tag and the interest words to obtain at least two clusters, and obtaining corresponding embedded points of word vectors contained in each cluster, which are corresponding to the clusters and the keywords in a preset embedded point set, as primary embedded points and secondary embedded points, generating dynamic embedded points based on the primary embedded points and the secondary embedded points, and carrying out information recommendation on the target object.

Inventors

LIU YANG
Xiong Huanwei

Assignees

深圳市东信时代信息技术有限公司
深圳市东信时代信息技术有限公司

Dates

Publication Date: 20260421
Application Date: 20210817
Priority Date: 20210817

Claims (9)

1. An information recommendation method, characterized in that the information recommendation method comprises: Acquiring registration data of a target object, and generating an initialization tag of the target object based on the registration data; Collecting behavior data of the target object, and determining interest words of the target object according to the behavior data; Clustering is carried out based on the initialization tag and the interest words to obtain at least two clustering clusters, and keywords corresponding to word vectors contained in each clustering cluster; Respectively obtaining corresponding buried points of the cluster and the keywords in a preset buried point set, and taking the buried points as a primary buried point and a secondary buried point; Generating a dynamic buried point based on the primary buried point and the secondary buried point, constructing a mapping relation between the primary buried point and the secondary buried point, setting the visual attribute of the primary buried point to be visual, setting the visual attribute of the secondary buried point to be invisible, taking the triggered primary buried point as a target primary buried point when any one of the primary buried points is triggered, acquiring the secondary buried point corresponding to the target primary buried point based on the mapping relation as a target secondary buried point, updating the visual attribute of the target secondary buried point to be visual, rendering and displaying the target secondary buried point at a client, and recommending information to the target object based on the dynamic buried point.
2. The information recommendation method of claim 1, wherein the determining the interest word of the target object according to the behavior data comprises: analyzing and filtering the behavior data to obtain user keywords; Training the user keywords in a word vector mode to obtain initial word vectors; And respectively calculating the Euclidean distance from the initial word vector to the word vector corresponding to each preset interest word aiming at each initial word vector, and taking the preset interest word corresponding to the Euclidean distance with the minimum value as the interest word of the target object.
3. The information recommendation method of claim 1, wherein the clustering process is performed based on the initialization tag and the interest word to obtain at least two clusters, and the keywords corresponding to the word vectors included in each cluster include: Converting the initialization tag into a word vector to obtain a first word vector, and converting the interest word into a word vector to obtain a second word vector; Acquiring preset weight information, and respectively carrying out weighting treatment on the first word vector and the second word vector based on the preset weight information to obtain an updated first word vector and an updated second word vector; And clustering the updated first word vector and the updated second word vector by adopting a K-Means aggregation algorithm to obtain at least two clustering clusters and keywords corresponding to the word vectors contained in each clustering cluster.
4. The information recommendation method of claim 1, wherein the respectively obtaining the corresponding buried points of the cluster and the keyword in the preset buried point set as the first-level buried point and the second-level buried point includes: Taking the corresponding buried points of each cluster in the preset buried point set as primary buried points, and taking the corresponding buried points of each keyword in the preset buried point set as initial secondary buried points; aiming at each primary buried point, generating a buried point information thermodynamic diagram by adopting an iframe mode based on the distance of keywords in a cluster corresponding to the primary buried point; and selecting the initial secondary buried point based on the buried point information thermodynamic diagram and a preset selection mode to obtain the secondary buried point.
5. An information recommendation device, characterized in that the information recommendation device comprises: the tag acquisition module is used for acquiring registration data of a target object and generating an initialization tag of the target object based on the registration data; the interest word determining module is used for collecting behavior data of the target object and determining interest words of the target object according to the behavior data; The clustering processing module is used for carrying out clustering processing based on the initialization tag and the interest word to obtain at least two clustering clusters and keywords corresponding to word vectors contained in each clustering cluster; The embedded point layering module is used for respectively acquiring embedded points corresponding to the cluster and the keywords in a preset embedded point set to serve as primary embedded points and secondary embedded points; The information recommendation module is used for generating dynamic buried points based on the primary buried points and the secondary buried points, constructing a mapping relation between the primary buried points and the secondary buried points, setting visual attributes of the primary buried points to be visual, setting visual attributes of the secondary buried points to be invisible, taking the triggered primary buried points as target primary buried points when any one of the primary buried points is triggered, acquiring the secondary buried points corresponding to the target primary buried points based on the mapping relation and taking the secondary buried points as target secondary buried points, updating the visual attributes of the target secondary buried points to be visual, enabling the target secondary buried points to be rendered and displayed at a client, and recommending information to the target object based on the dynamic buried points.
6. The information recommendation apparatus of claim 5, wherein the interest word determining module comprises: the data analysis unit is used for analyzing and filtering the behavior data to obtain user keywords; The word vector generation unit is used for training the user keywords in a word vector mode to obtain initial word vectors; the interest word determining unit is used for respectively calculating the Euclidean distance from the initial word vector to the word vector corresponding to each preset interest word according to each initial word vector, and taking the preset interest word corresponding to the Euclidean distance with the minimum value as the interest word of the target object.
7. The information recommendation apparatus of claim 5, wherein the The clustering processing module comprises: The word vector conversion unit is used for converting the initialization tag into a word vector to obtain a first word vector, and converting the interest word into the word vector to obtain a second word vector; the word vector updating unit is used for acquiring preset weight information, and respectively carrying out weighting processing on the first word vector and the second word vector based on the preset weight information to obtain an updated first word vector and an updated second word vector; The clustering unit is used for clustering the updated first word vector and the updated second word vector by adopting a K-Means aggregation algorithm to obtain at least two clustering clusters and keywords corresponding to the word vectors contained in each clustering cluster.
8. 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 information recommendation method according to any of claims 1 to 4 when executing the computer program.
9. A computer readable storage medium storing a computer program, characterized in that the computer program, when executed by a processor, implements the information recommendation method according to any one of claims 1 to 4.

Description

Information recommendation method, device, computer equipment and storage medium Technical Field The present invention relates to the field of data processing, and in particular, to an information recommendation method, apparatus, computer device, and medium. Background The Internet can provide abundant information resources for users, along with the rapid development of Internet technology, more and more users acquire information through the Internet, different people have different interesting target objects, and interest preference of the users needs to be acquired for accurately recommending information to the users. At present, the user interest preference is mainly determined by a buried point mode, namely counting the frequency of clicking each buried point application by a user, but the mode is limited by the number and distribution of the buried point applications, has a certain limitation, and when the number of the buried points is set to be large, a background is required to monitor a large number of buried point messages at the same time, so that more resources are occupied, and meanwhile, too many contents which are not interesting by the user are mixed in the information recommendation mode, so that the pertinence of information recommendation is not strong, and when the number of the buried points is small, preference information of the user is easy to miss, and the information recommendation is not accurate. Disclosure of Invention The embodiment of the invention provides an information recommendation method, an information recommendation device, computer equipment and a storage medium, so as to improve the accuracy of current information recommendation. In order to solve the above technical problems, an embodiment of the present application provides an information recommendation method, including: Acquiring registration data of a target object, and generating an initialization tag of the target object based on the registration data; Collecting behavior data of the target object, and determining interest words of the target object according to the behavior data; Clustering is carried out based on the initialization tag and the interest words to obtain at least two clustering clusters, and keywords corresponding to word vectors contained in each clustering cluster; Respectively obtaining corresponding buried points of the cluster and the keywords in a preset buried point set, and taking the buried points as a primary buried point and a secondary buried point; And generating a dynamic buried point based on the primary buried point and the secondary buried point, and recommending information to the target object based on the dynamic buried point. Optionally, the determining the interest word of the target object according to the behavior data includes: analyzing and filtering the behavior data to obtain user keywords; Training the user keywords in a word vector mode to obtain initial word vectors; And respectively calculating the Euclidean distance from the initial word vector to the word vector corresponding to each preset interest word aiming at each initial word vector, and taking the preset interest word corresponding to the Euclidean distance with the minimum value as the interest word of the target object. Optionally, the clustering processing is performed based on the initialization tag and the interest word to obtain at least two clusters, and the keywords corresponding to the word vectors contained in each cluster include: Converting the initialization tag into a word vector to obtain a first word vector, and converting the interest word into a word vector to obtain a second word vector; Acquiring preset weight information, and respectively carrying out weighting treatment on the first word vector and the second word vector based on the preset weight information to obtain an updated first word vector and an updated second word vector; And clustering the updated first word vector and the updated second word vector by adopting a K-Means aggregation algorithm to obtain at least two clustering clusters and keywords corresponding to the word vectors contained in each clustering cluster. Optionally, the obtaining the corresponding buried points of the cluster and the keyword in the preset buried point set respectively includes: Taking the corresponding buried points of each cluster in the preset buried point set as primary buried points, and taking the corresponding buried points of each keyword in the preset buried point set as initial secondary buried points; aiming at each primary buried point, generating a buried point information thermodynamic diagram by adopting an iframe mode based on the distance of keywords in a cluster corresponding to the primary buried point; and selecting the initial secondary buried point based on the buried point information thermodynamic diagram and a preset selection mode to obtain the secondary buried point. Optionally, the generating the dynamic buried po