Music Personalized Recommendation System Based on Hybrid Filtration

Duet to the diversity and fuzziness of music itself and the accuracy of music melody, the recommendation algorithm with highest accuracy can not match users' test completely. For such problem, this article proposes a hybrid recommendation algorithm based on collaborative filtering algorithm and music genes, and design a personalized music recommendation system. The system first computes recommendation results according to the collaborative filtering algorithm, and discover the potential interests of the user. Then each music is weighted by preference on the genes of collected music. After weight selection the music which has prior preference is taken as recommendation results. Finally two recommendation results are performed weighted combination and filtering to make recommendation. The experimental data indicate the improved scheme can increase the accuracy of recommendation and meet users' demand from different levels.