基于用户行为数据的非负矩阵分解音乐软件推荐算法研究

打开文本图片集

中图分类号：TP391.4 文献标识码：A 文章编号：2096-4706（2025）08-0111-06

Abstract： With the popularity of intermet music services，how to accurately recommend music for users has become an importantresearch topic.This paperaimsattheshortcomingsof theexistingmusicrecommendationsystemindealingwith problemssuchascold-startanddatasparsity.AmusicrecommendationalgorithmbasedonNon-NegativeMatrixFactorization （NMF） is proposedThe studyusesadataset fromacolaborationproject with NetEaseCloud Music，whichcontains more than 57 millon music interactionrecordsofmore than2millonusers.Byintroducinguserbehavior weightsandsparseconstraints， weighted NMFand sparse NMF models are constructed respectively.The experimental results show that the weighted NMF performs best when dealing with high-frequency interactive users，and the F1 score reaches .The sparse NMF has more advantages in dealing withcold-start users.Forusers with fewer than1O interactions，therecommendation accuracy is 1 5 % higher than that of the basic NMF.The research results provide new solutions for the optimization of the music recommendation system.

Keywords：Machine Learning; music recommendation model; NMl

0 引言

随着互联网和信息技术的快速发展，数字音乐产业得到了迅猛扩展。（剩余8976字）