Title: Research on long- and short-term music preference recommendation method integrating music emotional attention

Authors: Yan Yang

Addresses: Department of Music and Dance, Hunan University of Science and Engineering, Yongzhou, 425199, China

Abstract: In order to improve the effect of user music personalised recommendation, a hybrid music personalised recommendation model based on attention mechanism and multi-layer LSTM is proposed from the perspective of user music emotion and behaviour data. Using multi-layer LSTM to mine users' long-term and short-term music preferences, the model can analyse users' music emotional attributes in combination with attention mechanism. The research results show that the recommendation accuracy of the AM-LSTPM model is 97.86%, the recall rate is 98.91%, and the NDCG@10 values of the model on the two datasets are 0.5771 and 0.5437, which can effectively provide users with targeted personalised music recommendation services. The research, based on the modelling of users' long-term and short-term music preferences and integrating users' music emotional attention analysis, provide users with high-quality targeted music recommendation services, and have important value in promoting the improvement of music streaming media service quality.

Keywords: long short-term memory; LSTM; attention mechanism; music; personalised recommendation; emotion.

DOI: 10.1504/IJNVO.2023.133873

International Journal of Networking and Virtual Organisations, 2023 Vol.28 No.2/3/4, pp.381 - 397

Received: 29 Dec 2022
Accepted: 12 Jun 2023
Published online: 04 Oct 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article