An Application Research Of Music Recommendation Based On FP-growth Algorithm

Personalized recommendation has applications in many aspects of social life, and music recommendation is an important part of them. The growth of digital music is exponentially currently.When users want to search and found their favorite music from the mass music, they often cannot reach their own purpose. In the face of such demand, the domestic and foreign scholars try to study music retrieval problems. Most common music retrieval methods are depended on the existing libraries. What’s worse, only when the user remember all kinds of information of the target library can music be retrieved later. In the case of the ever-accelerating pace of social life, this approach is clearly not in line with users’ requirements. What users need is fast response, so the system requires a clear understanding of the needs of users, to understand the users’ interest, and play out the corresponding music, then achieve results that users want. Practices show that music recommendation system is able to meet the demand of the audience and can find interesting songs for users. At present, it is urgent to recommend the most unique play list in uses’ interests for different users.Music recommendation technology is some methods that based on labels, content, collaborative filtering and so on. The research of recommendation system based on association rules algorithm is still in its infancy, and the application research on music is also rare, so on the basis of reviewing the data mining technology and music recommendation technologies, we compared and analyzed FP-growth with other music recommendation algorithms,as to say,collaborative filtering recommendation algorithm and item-based recommendation algorithm, it was proved that the music recommendation depended on the algorithm described in this paper had its advantage and effect on solving recommendation problem. Then we applied FP-growth of association rules algorithm to data mining and recommendation of music. At the same time, the music database had been transformed effectively which could improve the mining efficiency and accuracy. After the subjective test of users’ satisfaction in the result of the music recommendation, it showed that the music recommendation based on the algorithm described here did have its feasibility and practical value. Finally, the actual research work is summarized and we give an outlook of the improvement of algorithms and music recommendation technology.The technology of network music recommendation system remains to be development. This study makes an attempt to the application of FP-growth algorithm in music, and offers technology support to access of mass music resources which can be utilized efficiently.