Research on intelligent music personalized recommendation algorithm based on MLP-Mixer efficient feature extraction

With the rapid development of the digital music market, users’ demand for personalized music recommendations is increasing daily. Traditional recommendation algorithms often face problems such as insufficient feature extraction and low recommendation accuracy when dealing with large-scale and high-dimensional music data. Therefore, this study proposes an intelligent music personalized recommendation algorithm based on efficient feature extraction of MLP-Mixer. The algorithm uses MLP-Mixer’s multi-layer perceptron network to effectively capture deep-seated features in music data, thereby improving the accuracy and diversity of the recommendation system. In the experimental session, we constructed a dataset containing 100,000 songs and 1 million user ratings, using mean square error (MSE) and Accuracy as evaluation indicators. The experimental results show that compared with the traditional recommendation algorithm, the proposed algorithm reduces the MSE index by 15% and improves the Accuracy index by 8%. In addition, the user satisfaction survey shows that the user satisfaction of the recommendation system using the new algorithm has increased by 12%, significantly enhancing the user experience. This study not only verifies the effectiveness of MLP-Mixer in music feature extraction but also provides a new technical path for intelligent music personalized recommendation.