Title: Study on improved personalised music recommendation method based on label information and recurrent neural network

Authors: Yali Zhang

Addresses: Music Department, Henan Polytechnic, Zhengzhou 450046, Henan, China

Abstract: In order to improve the problems of low accuracy and the time-consuming period of traditional personalised music recommendation methods, a personalised music recommendation method based on label information and a recurrent neural network is proposed in this paper. Firstly, the music label information is extracted, and the music label information is clustered according to the label similarity. Secondly, the music label information clustering results are decomposed by tensor, and all tensor decomposition data are fused to generate the target user recommendation list. Finally, the recurrent neural network is used to select personalised music from the user recommendation list and recommend it to users. The simulation results show that the accuracy of personalised music recommendation is always more than 93%, and the recommendation time is always less than 0.6 s.

Keywords: label information; recurrent neural network; individualisation; music recommendation; similarity calculation; tensor decomposition.

DOI: 10.1504/IJICT.2024.135328

International Journal of Information and Communication Technology, 2024 Vol.24 No.1, pp.48 - 59

Received: 21 Oct 2021
Accepted: 06 Dec 2021
Published online: 05 Dec 2023 *