Social recommendation method based on multi-feature heterogeneous graph neural network
What is Al technical title?
Al technical title is built by PatSnap Al team. It summarizes the technical point description of the patent document.
A neural network and recommendation method technology, applied in the field of data mining information recommendation, can solve problems such as information loss, single focus on user interaction or user interest topics, and no consideration of the impact of the recommendation system, to achieve the effect of improving accuracy and user experience
Active Publication Date: 2021-08-13
JINAN UNIVERSITY
View PDF11 Cites 7 Cited by
Summary
Abstract
Description
Claims
Application Information
AI Technical Summary
This helps you quickly interpret patents by identifying the three key elements:
Problems solved by technology
Method used
Benefits of technology
Problems solved by technology
[0003] In the existing social recommendation system technology, most of them only consider some sparse information (such as interaction between users, comments, likes, etc.), and do not consider the influence of multiple factors on the recommendation system
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more
Image
Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
Click on the blue label to locate the original text in one second.
Reading with bidirectional positioning of images and text.
Smart Image
Examples
Experimental program
Comparison scheme
Effect test
Embodiment 1
[0058] Such as figure 1 As shown, this embodiment provides a social recommendation method based on a multi-feature heterogeneous graph neural network, including the following steps:
[0059] S1: Extract and preprocess various attribute information of users and topics
[0060] At the same time, various attribute information of users (such as nickname, age, city, etc.) and various attribute information of topics (such as subject, popularity, profile information, etc.) are extracted. The semantic information related to natural language is encoded by embedding feature using word2vec method. , and the rest can be one-hot encoded with discrete-valued data.
[0061] This embodiment extracts and preprocesses various attribute information of users and topics, specifically including the following sub-steps:
[0062] S1.1: Multi-feature extraction
[0063] At the same time, it initially extracts various attribute information of social platform users (such as nickname, age, city, etc.)...
Embodiment 2
[0106] This embodiment provides a social recommendation system based on a multi-feature heterogeneous graph neural network, including: an information extraction module, an encoding preprocessing module, an initial feature vector output module, a heterogeneous graph construction module, and a heterogeneous graph neural network model construction module , an attention score calculation module, a comprehensive attention score calculation module, an information aggregation and update module, a feature vector similarity calculation module and a recommendation result output module;
[0107] In this embodiment, the information extraction module is used to simultaneously extract multiple attribute information of users and multiple attribute information of topics;
[0108] In this embodiment, the encoding preprocessing module is used to perform encoding preprocessing on various attribute information of users and various attribute information of topics;
[0109] In this embodiment, the ...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more
PUM
Login to view more
Abstract
The invention discloses a social recommendation method based on a multi-feature heterogeneous graph neural network. The method comprises the following steps: extracting various attribute information of social network users and topics for coding; processing the user coding information and the topic coding information through a multi-layer perceptron to obtain initial feature vector representation of each user and topic; establishing a heterogeneous graph by taking users and topics as nodes, inputting the heterogeneous graph into a heterogeneous graph neural network, performing information transmission in the graph in combination with an attention mechanism, and updating feature vector representation; and performing similarity calculation on the user feature vectors, and selecting top-k users and top-k topics with the highest similarity with the user vectors for recommendation. Various types of attribute information of the users and the topics is extracted at the same time, the users and the topics serve as nodes at the same time to establish the heterogeneous graph, social information can be mined more comprehensively, information transmission and aggregation are carried out through the heterogeneous graph neural network, features of the users and interested topics of the users are deeply fused, and recommendation accuracy and user experience are improved.
Description
technical field [0001] The invention relates to the technical field of data mining information recommendation, in particular to a social recommendation method based on a multi-feature heterogeneous graph neural network. Background technique [0002] In recent years, with the rapid development of the Internet and information technology and its subsidiary industries, Internet services and transactions have become more and more popular, and people are more and more inclined to use the Internet for social activities. A large amount of user behavior information, such as mutual attention, comments, likes, and exchanges on topics of common interest among users, has caused information overload. Although this provides users with rich information, it inevitably makes users spend a lot of time on a large amount of information. information that they are not interested in, so information overload is both a challenge and an opportunity for users and social platforms. The emergence of the...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more
Application Information
Patent Timeline
Application Date:The date an application was filed.
Publication Date:The date a patent or application was officially published.
First Publication Date:The earliest publication date of a patent with the same application number.
Issue Date:Publication date of the patent grant document.
PCT Entry Date:The Entry date of PCT National Phase.
Estimated Expiry Date:The statutory expiry date of a patent right according to the Patent Law, and it is the longest term of protection that the patent right can achieve without the termination of the patent right due to other reasons(Term extension factor has been taken into account ).
Invalid Date:Actual expiry date is based on effective date or publication date of legal transaction data of invalid patent.
Login to view more 
Login to view more