Matrix decomposition recommendation method in graph construction framework

Abstract

The invention discloses a matrix decomposition recommendation method in a graph construction framework. The method comprises the following steps of (1) establishing a score matrix according to user score data; (2) adding constraint terms to a manifold regularization-based score matrix decomposition model to update similarity information so as to build a new decomposition model; (3) calculating initial similarity between articles based on a score, tag or topic model; (4) randomly initializing to-be-solved latent semantic eigenvectors in the decomposition model and offset vectors of the articles and a user; (5) based on the initial similarity, adaptively updating parameters in the decomposition model by using a stochastic gradient descent method; and (6) obtaining a complete final score matrix, and providing recommendation for the user according to the score matrix. According to the method, the deficiency of a recommendation algorithm depending on similarity in inaccurate similarity calculation can be made up for.

Description

technical field [0001] The invention relates to the field of recommendation learning, in particular to a matrix decomposition recommendation method under a graph construction framework. Background technique [0002] At present, existing recommendation technologies include content-based recommendation, collaborative filtering recommendation, and so on. The content-based recommendation algorithm constructs user preference information based on some historical information of users, calculates the similarity between recommended items and user preferences, and recommends items with high similarity to users. Collaborative filtering algorithms can be divided into neighborhood-based collaborative filtering algorithms and matrix factorization algorithms. Collaborative filtering recommendation algorithms based on neighbors include user-based collaborative filtering recommendation algorithms and item-based collaborative filtering algorithms. Matrix decomposition is to complete the sco...

AI Technical Summary

Problems solved by technology

[0008] Purpose of the invention: The present invention aims at the problems existing in the prior art, and provides a new matrix decomposition recommendation method that combines manifold regularization and neighbor similarity improvement. This method takes the similarity matrix as a parameter and continuously corrects iterative optimization, which can compensate Insufficiency of similarity-dependent recommendation algorithms when similarity calculations are imprecise

Images