The Design And Implementation Of A Hybrid Music Recommendation System Based On Music Gene

Music is a sound art which can express people’s thoughts,feelings and social life,and is also the most impressive forms of the art.People create music,record music,and enjoy music. The rapid development of information technology brings the expanding of the amount of the information.This makes us easier to get more network resource, but at the same time, this makes us more confused about filtrating the information.For the music,with the increase pursuit of the sprit and entertainment,the number of the music is growing fast,which makes us more hard to retrieve and find the music we like. The researchers all around the world are putting their efforts on the research about music retrieval.Traditional music retrieval mostly uses the music information.this method need users to remember the accurate music information,and we cannot judge that if we will like some new music if we know nothing about the information of this music.In face of the rapid living environment,most people need the background music,they hope it easier to find the music and the music is also in accordance with their interests.By this way,people can meet the music they like by accident whatever the condition they are in.This is also the primary task of the music recommendation system.In consideration of the diversity and fuzziness when people understand and love the music, the accuracy of the melody of the music, Even the recommendation results of the most accurate algorithm can not match the users’ interests perfectly.In this article,the primary job is to present a hybrid recommendation algorithm which combines the collaborative filtering recommendation algorithm and the recommendation algorithm based on music gene, design and develop a music recommendation system based on this hybrid recommendation algorithm.In this system, the first step is to calculate and present a recommendation results by collaborative filtering recommendation algorithm.This algorithm can make up the deficiency of the recommendation by the analysis of the content to some extent,and find the potential hobbies and interests of the uses which they even don’t know,and this algorithm is also easier to realize.The recommendation algorithm based on music gene has good scalability.The method used in this system to calculate the similarity of the music gene is different from the traditional method.This system calculates the preference degree of every music gene by listening records of the users,adds a weight for the music by the preference degree,and chooses the music which has a relatively higher weight as the recommendation results.Finally the system gives a weight to both of the recommendation results’ and filters the hybrid results.The music recommendation system of this article was developed by traditional Software engineering development process,In the Interface design,some new buttons were added to increase the users’ necessary interaction,and improve the cognitive ability of the system.Through the result evaluation of the experiment,the hybrid recommendation algorithm proposed in this paper can improve the effect of the recommendation of recommendation system to some extent.