基于领域覆盖算法的音乐情感识别

Music Emotion Recognition Based on Neighborhood Covering Algorithm

音乐中具备很多情感的信息。文中通过分析音乐特征并用领域覆盖算法对音乐情感分类进行研究。音乐情感分类主要包括两个阶段:特征提取和分类。首先,通过Matlab语言提取音乐的特征,将提取到的特征值构建训练样本,然后使用训练样本训练领域覆盖算法分类器,得到音乐情感分类器,从而实现音乐的情感自动分类。文中借鉴Weiner、Graham的情感分类方法,将音乐分为开心和悲伤两类,并尝试用多种不同的音乐特征组合训练领域覆盖分类器,分析基于领域覆盖算法的音乐情感识别效果。

Music is strongly associated with emotions. In this paper,neighborhood covering algorithm is used to analyze musical genre so that music can be classified into different categories. Musical genre classification task falls into two major stages：feature extraction and classification. First of all,Matlab is used to extract the characteristics of the music,the characteristic values are applied to construct the training sample,then adopt the training sample to train the neighborhood covering algorithm classifier,obtaining the music emotion classi-fier and realizing the emotional automatic classification of music. According to emotion classification method of Weiner and Graham,di-vide the Chinese popular music into two categories：happy and sad,then try to use different music feature combinations to train neighbor-hood covering algorithm classifier as a method for music emotion recognition.