计算机技术与发展Issue(8)：184-187,193,5.DOI:10.3969/j.issn.1673-629X.2015.08.039
基于音乐内容和歌词的音乐情感分类研究
Research on Music Emotion Classification Based on Music Content and Lyrics
摘要
Abstract
According to the music emotion classification,an approach of multi-modal music emotion category combining music content and lyrics is proposed to compensate for lack of the single modal music emotion classification method that only uses music content for classification. Mainly discuss how to use lyrics for music emotion classification and combine music lyrics and content to improve the clas-sification accuracy. Using feature selection algorithm based on CHI and quadratic dimension reduction method based on Latent Semantic Analysis ( LSA) effectively improves the efficiency of text classification. For multi-modal fusion problem,propose an improved LFSM fusion method based on the traditional LFSM fusion method,and verify its feasibility through some experiments and compare the im-proved LFSM fusion method with the others. The results show that the accuracy of the improved method is highest,reaching 79. 51%, that verify the effectiveness of the method.关键词
音乐情感分类/CHI特征选择/潜在语义分析/多模态融合Key words
music emotion classification/CHI feature selection/LSA/multi-modal fusion分类
信息技术与安全科学引用本文复制引用
邵曦,陶凯云..基于音乐内容和歌词的音乐情感分类研究[J].计算机技术与发展,2015,(8):184-187,193,5.基金项目
国家自然科学基金资助项目(60902065) （60902065）
