Design Of Music Recommendation Algorithm For Mobile Devices

With the rapid development of digital music, music information overload problems have become increasingly prominent, personalized music recommendation technology as an effective means to solve the problem of information overload of music, received extensive attention and research. Traditional personalized music recommendation system, according to the user’s historical data, use different personalized recommendation algorithm to provide users with personalized service. But in mobile device applications, in particular on-board equipment such as wireless phone signal faster or bad times, or switch music on demand are often slow to respond, the server unable to respond to the changing needs of users, and because the server-side push, cause buffer music files, which are buffered in the degree of user’s preference does not match, will result in unnecessary waste of traffic.Based on this, the personalized recommendation method extends to the mobile device side, according to the user record mobile device ends with feedback recorded on the mobile device side of the music group and the calculated user’s preference degrees for the user to make personalized music recommendations feedback data while moving constantly receiving end user to listen to music in different musical groups in the time to design a dynamic adjustment algorithm to adjust the degree of match weights of each category of music in real time, according to the size of the weights of different groups within music files modest prefetch, allowing users to switch songs and groups can reduce the waiting time, smooth listen to music, and save network traffic.We completed the design and implementation of APP-based music paper proposed personalized music recommendation method, the integrated music APP Ltd. in Wuhan Sunshine Road passenger automotive products in pony blah. Through comparative analysis of test users to listen to feedback data, use of this method include music recommendation music recommendation application system to reduce waiting time, save data traffic aspects of performance better, to enhance hit rate recommendation system by 11%, confirming the method effectiveness.Music group since the initial weights of the method used is an expert score, in fact, in most cases is not consistent with the preferences of the user, it will cause weight in the initial phase of the slow convergence problem. On the other hand, people’s tastes in music vary from place will continue to change, how to analyze both long-term tracking of user preferences, but also gives fresh recommendation is appropriate future needs further study.