A recommendation method based on a graph convolutional neural network

Abstract

The application provides a recommendation method based on a graph convolutional neural network, comprising the following steps: constructing an adjacency matrix A based on a user-item historical interaction matrix R; constructing a residual module with initial feature connection to capture high-dimensional collaborative signals; projecting original visual and text features into the same vector space; filtering out redundant information through a nonlinear activation function under the guidance of ID features; constructing a modal-aware item-item relationship graph based on KNN sparsification technology to enhance the modal features of items; introducing a behavior-guided multi-modal feature mining module for extracting multi-modal collaborative signals; introducing a hyperparameter gamma to fuse image features E v and text features E t , predict the items that the user u is most likely to like, and recommend the top N items with the highest scores to the user; performing self-supervised learning in the whole stage, obtaining a whole-stage self-supervised learning loss function, and optimizing the whole-stage self-supervised learning loss function based on a Bayesian personalized ranking loss.

Images