科研之友
微信
新浪微博
Facebook
分享链接
摘要
Taking user’s emotion in music retrieval and recommendation as application background, this paper presents a method to extract the features associated with music emotion from the existing physical features. Focusing on degree of stability, differentiation and relativity, the method simplifies the existing multidimensional features and extracts the music physical features associated with emotion. We verify the presented method on data from baidu.com. The experimental results show that the proposed method can improve the classification efficiency about ten percent under the condition that the precision can be guaranteed.
