Research And Implementation On Personalization Music Recommendation Method Oriented User's Music Preferences

The network has gradually become a major channel for the public to obtain information and consumption,the rapid expansion of information leads to extreme redundancy of network data,which makes people not able to find their focus quickly and accurately.Music,one of the traditional entertainments,has the same problem of information overload.If the music website was built in the structure where music products were simply listed one by one,users had to spend a lot of time to browse through massive irrelevant information before they can find out their own interests.During the browsing process,users may lose interests and leave the website.Therefore,according to the user's historical track record,finding out the users' preferences and providing recommendations of the music they may like will greatly reduce the user's manual search time,optimize the user experience during browsing the site,fundamentally improve the user's interest and attention degree of the site so as to achieve to improve the quality of service and increase the value of the website.In this thesis,an in-depth study on personalized music recommendation method oriented users' users' preferences was conducted,using labels to sort the music,the efficiency and accuracy of the recommendation can be improved by analyzing the music preferences of users.In this thesis from the perspective of software engineering,the development and implementation of personalized music recommendation system was described in detail.The main work includes the following three aspects:First,questions were raised based on the research background and research status both at home and abroad,determine the significance of the research.The steps of the solution for the personalized music recommendation oriented users music preferences' based on previous material: firstly MapReduce was used to streamline the data set,then Apriori algorithm was used to calculate the association rules between songs,finally the sliding window technology and the Ebbinghaus forgetting curve were used to construct the user interest model to realize personalized recommendation.Second,conduct requirement analysis,construct the system structure according to the requirement analysis,determine the function modules of the system,design the system in detail,develop the "HEART Music FM" personalized music recommendation system.Third,program and test the system according to the design,ensure that the system can be presented fully and independently,the recommended method can successfully achieve the desired results,and ultimately shield out the useless information,simply present the users' preferences and potential preferences of the music in the site for the user to choose.