A CF recommendation method fusing matrix decomposition and user project information mining

Abstract

The invention provides a CF (collaborative filtering) recommendation method fusing matrix decomposition and user project information mining. The CF recommendation method comprises the following steps:reading historical score data and project type data information of a user on an article; based on the FunkSVD model, optimizing and decomposing the user score matrix, and adding a similarity factor to calculate and generate a user score prediction matrix; calculating optimal similarity by optimizing CF users and project information occupying different proportions, predicting user scores, and generating Top-N recommendation lists. The method has the advantages that (1) the user scoring matrix is optimized and decomposed based on the FunkSVD model, and the trust factor is added to predict the user scoring matrix, so that the problem of low prediction accuracy caused by data sparseness of a traditional matrix decomposition model is relieved; (2) similarity is calculated based on the user information and the project information, and the problem of cold start caused by excessive dependence on historical data in a traditional recommendation algorithm is solved; and (3) a trust degree relationship between users is introduced, so that the recommendation precision and interpretability of a traditional CF recommendation algorithm are improved.

Description

technical field [0001] The invention belongs to the technical field of movie recommendation, and relates to a CF movie recommendation method integrating matrix decomposition and mining user item information. Background technique [0002] In the era of big data intelligence, the personalized recommendation system can recommend content of interest to users from massive data resources by studying users' preferences, and provide users with more accurate information recommendation services by learning and collecting user information. It allows users to obtain the information they want at a lower cost, and is used to help customers obtain resources that meet their interests, preferences and needs from a large amount of data information on the Internet. [0003] In the field of movie recommendation, collaborative filtering recommendation technology is one of the most successful technologies in the current recommendation system, which can be divided into two categories: user-based c...

AI Technical Summary

Problems solved by technology

[0005] Personally recommend movies that they may be interested in to different users, overcome the problems of data sparsity, cold start and weak scalability in traditional algorithms, and improve the accuracy of recommendation algorithms to achieve customized recommendations

Images