Recommendation method fusing content awareness and feature similarity

Abstract

A recommendation method fusing content awareness and feature similarity relates to the cross technical field of recommendation systems and distributed computing, and comprises the following steps: firstly, decomposing playing frequency data by using a weighted matrix decomposition model: decomposing the playing frequency data into a product of a transpose matrix WT of a user preference matrix and an item (namely music) attribute matrix H; introducing content information into weighted matrix decomposition in consideration of the characteristics of a project (music): adopting a mode of rewriting a project attribute matrix H, specifically, rewriting the project attribute matrix H by adopting a content characteristic zi to obtain a rewritten project attribute matrix Hi; obtaining a prediction matrix A (WT * Hi) by adopting the product of a transpose matrix WT of the user preference matrix and the rewritten project attribute matrix Hi; performing cosine similarity calculation on the timbre and pitch of the item (music) to obtain a prediction matrix B; and fusing the obtained prediction matrix A and the prediction matrix B by using a fusion factor to obtain a recommended value F (mu, i), of which the specific expression is F (mu, i) = theta A (mu, i) + (1-theta) B (mu, i). According to the invention, the cold start problem in the recommendation process can be effectively solved.

Description

technical field [0001] The invention relates to the cross technical field of recommendation system and distributed computing, in particular to a recommendation method integrating content perception and feature similarity. Background technique [0002] The current mainstream music recommendation system is based on collaborative filtering, which predicts the user's interest based on the user's listening habits and similarity with other user profiles. This type of method is unknown to the content of the song, and cannot recommend new songs without historical records, and there is a technical problem of cold start. In order to solve the above technical problem, the present invention provides a recommendation method that combines content perception and feature similarity. Contents of the invention [0003] The purpose of the present invention is to provide a recommendation method that integrates content perception and feature similarity in order to solve the technical problems ...

AI Technical Summary

Problems solved by technology

This type of method is unknown to the content of the song, and cannot recommend new songs without historical records, and there is a technical problem of cold start. In order to solve the above technical problems, the present invention provides a recommendation method that combines content perception and feature similarity

Images