摘要
音乐作为人类情感表达和文化交流的重要媒介,其情感分析越来越多的受到学者专家的重视。针对歌词文本进行情感倾向性分析,提出一个构建音乐情感词典的模型。首先,对指定网站及类别标签的歌词文本进行收集与预处理;其次,在手动整理的歌词基准词汇库的基础上,使用基于点间互信息法测量相似度的方法对情感词典进行扩充;最后在该情感词典的基础上进行歌曲情感分类,实验结果准确率均在70%以上,达到较为理想的结果。
Music, as an important medium of human emotional express and cultural exchange, has attracted more and more attention from scholars and experts. According to sentiment orientation analysis of lyrics text, puts forward a model of music emotion dictionary. Firstly, collects and preprocesses the lyrics text of the designated website and category label. Secondly, on the basis of the manual lexicon base of the lyrics, uses the method of measuring similarity based PMI to expand the emotional dictionary. Finally, the emotional classification of the songs is carried out based on the emotional dictionary. The accuracy rate of the experiment is over 70%, which is idealized.
作者
智昕
周日贵
ZHI Xin;ZHOU Ri-gui(College of Information Engineering,Shanghai Maritime University,Shanghai 201306)
出处
《现代计算机》
2018年第14期50-53,共4页
Modern Computer