Research Of Music Recommendation And Visualization Based On User Information And Music Characteristics

Music analysis is a popular topic in modern life.There are several aspects in music analysis.It contains music recommendation,music visualization and research on musical emotions and so on.There are thousands of information without clear classification on the Internet.Recommendation algorithm which is based on data mining emerges under that situation.As we know,connections between music pieces and users exists on the Internet.Music recommendation is accomplished by recommending music pieces for users based on some hidden data between songs and users.Music visualization is also an important part of music analysis.Normally speaking,we can listen to music.Music visualization is a kind of novel amusement style on a creative and innovative basis.Music visualization can be implemented by displaying some music characteristics as certain styles.During this process,we distract certain intrinsic characteristics of music pi eces through signal processing in frequency domain.Music recommendation and music visualization is realized through algorithm analyzing and project implementation separately in this paper.In Algorithm analyzing,a novel hybrid algorithm is proposed in this procedure.By adding an impact factor,the traditional collaborative algorithm is optimized.After that,an experimental research is performed and recommendation efficiency is valued by Mean Absolute Error.Influence of the number of neighboring and percentage of testing datasets have been verified separately during the experiments.All of those methods are included.The influence of impact factor of the hybrid method has also been verified in this paper.The results show that this novel algorithm can improve recommendation efficiency.In project implementation,a program application is realized based on Android platform.It can display visual view which is related to corresponding music when certain music is playing.Music analysis is realized by algorithm and project separately.The realization of music recommendation is more objective.At the evaluation stage,the effciency is evaluated by the Mean Absolute Error which is more persuative.Wherever the project program is more visible.People can evaluate the visualization effect through the visualization image of certain music pieces on the screen.