Design And Implementation Of Music System Based On Collaborative Filtering

The advent of social networks has led to the development of social music,which is the main way to share in real time on social networks.Due to the music sharing of the majority of users,there are a large number of music resources on the network.In addition,with the advent of the era of digital music,users' demands for resources such as apps have also been continuously improved,and the demand for music by users has become more diverse.For many users,80% of the massive music resources are noise.Therefore,how to quickly obtain music that may be of interest to users from a large amount of music information is a problem that the music industry must solve.However,the traditional music system focuses on music and user management.When users find music they like,they often spend a lot of time,which greatly reduces the user experience.Therefore,this paper plans to apply the collaborative filtering algorithm to the music system to recommend customized music for different users to meet the individual needs of users.Faced with a large amount of data,customized recommendations for users face cold start,inaccurate recommendation,slow recommendation,etc.Therefore,based on different scenarios,this paper studies and implements a hybrid collaborative filtering based on different users' needs.Recommended algorithms to achieve exclusive recommendations for system users.The system mainly collects the behavior records of the user's song playing,downloading,and collecting by implicitly,and uses the collaborative filtering recommendation algorithm based on the nearest neighbor user to recommend the song for the system user;for the song with the lyric information(in English),The word embedding of the text network is used to calculate the similarity between the songs,and then similar songs are recommended according to the user's historical records;in addition,the system records the user's access behavior,and the user is recommended and the historical song label through the label-based collaborative filtering algorithm.Similar songs.Finally,this paper demonstrates the music system implemented in this paper,which mainly includes the display of related core functions such as music recommendation and music review.The final experimental results show that the music system implemented in this paper not only has user registration,user login and other authentication modules,but also basic music management functions such as music playback and music collection.In addition,the system finally realizes the core recommendation function.The recommendation module can recommend suitable songs for different users based on the user's history,and the hybrid collaborative filtering recommendation algorithm is deduced by publicity and has credibility.Therefore,the test shows that this paper implements a powerful recommendation system based on the recommended algorithm of this paper,and has a wide range of application scenarios.