Session recommendation method based on graph neural network

Abstract

The invention discloses a session recommendation method based on a graph neural network. The method comprises the following steps: constructing a global session graph by adopting all session sequences; constructing a local session graph by adopting the current session sequence; inputting the global session graph into a global session graph neural network to obtain global information feature representation; inputting the local session graph into a local session graph neural network to obtain local information feature representation; integrating the global information feature representation and the local information feature representation to obtain session sequence feature representation; obtaining a final feature representation by adopting an attention mechanism; and calculating to obtain scores of the user and each commodity and outputting a final session recommendation result. According to the method, brand-new and different graph construction methods are used, information loss can be effectively reduced, meanwhile, global session information is added, collaborative information can be effectively added, and the recommendation effect and diversity of a current session are improved; the method is high in reliability, good in accuracy, scientific and effective.

Description

technical field [0001] The invention belongs to the technical field of computers, and in particular relates to a conversation recommendation method based on a graph neural network. Background technique [0002] With the development of economy and technology, various fields such as e-commerce, social networking, and long and short videos have developed rapidly. With the increase in the number of users, the popularization and application of big data technology, and the increasing requirements of users for user experience, it is particularly important to make personalized recommendations for different users. [0003] A session refers to a product sequence that users interact with over a period of time, such as click records on an e-commerce platform for a period of time or viewing records on a short video platform for a period of time. Sessions are usually separated according to time. The interaction information in a session can better reflect the user's interests and preferen...

AI Technical Summary

Problems solved by technology

However, the current session recommendation method based on the graph neural network model has the problem of simple construction of the session graph, which will lead to the lack of session information, so that the feature aggregation and extraction of the graph neural network model cannot be better utilized; in addition, in anonymous session recommendation In the method, the existing recommendation method cannot effectively use the global session information, and the effectiveness, diversity and accuracy of the recommendation need to be improved

Images