Song recommendation method based on hybrid recommendation algorithm

Abstract

The invention claims a song recommendation method based on a hybrid recommendation algorithm. The method includes the following steps: a user giving out a song play request, acquiring a user's song play log, and obtaining a user-song scoring matrix, determining from a target attribute item of the song play request from a preset user attribute item set, performing preliminary recommendation of songs by using the established user-song scoring matrix; calculating user preference degree of each song in the song play log according to the song play log; determining a recommendation list, and recommending the recommendation list to users; if half of the recommendation list is the same with an initial recommendation result, using the recommendation, if not, selecting the song with the second highest score as a target song, repeating the above steps to fill, until more than half is the same, recommending the song list to the user. The hybrid recommendation method can improve accuracy of song recommendation.

Description

technical field [0001] The invention belongs to the technical field of recommendation, in particular to a song recommendation method based on a hybrid recommendation algorithm. Background technique [0002] The traditional method of recommending songs is to obtain the type, album, and singer of the song that the user listens to, and recommend the corresponding song type, album, or singer to the user. However, most current music recommendation methods are based on collaborative filtering. The method has the following disadvantages: the music recommendation method implemented by the collaborative filtering algorithm cannot reflect the user's preference changes in time, and the update of the song correlation is slow, and the song correlation is easily affected by the noise data. Contents of the invention [0003] The present invention aims to solve the above problems of the prior art. A song recommendation method based on hybrid recommendation algorithm is proposed to improv...

AI Technical Summary

Problems solved by technology

[0002] The traditional method of recommending songs is to obtain the type, album, and singer of the song that the user listens to, and recommend the corresponding song type, album, or singer to the user. However, most current music recommendation methods are based on collaborative filtering. The method has the following disadvantages. The music recommendation method implemented by the collaborative filtering algorithm cannot reflect the user's preference changes in a timely manner, and the update of the song correlation is slow, and the song correlation is easily affected by noise data.

Images