摘要
情感计算是目前人工智能领域的热门课题,而词汇的情感计算又是准确完成文本情感计算的基础.目前情感词汇的获取大多采用人工获取的方法,如何自动地获取情感词汇,已成为当前情感计算研究亟需解决的问题.提出了情感词汇的自动提取机制,首先分析了情感词汇的一般语法规律,例如,重叠的规律,受否定词、程度副词修饰的规律等.然后在情感词汇的这些语法规律的基础上,运用CRF模型实现了情感词汇的自动获取.最后,分析了不同的语法规律对情感词汇自动获取的作用大小,并对实验结果进行了详细分析,实验结果表明情感词汇自动获取方法是有效的.
Affective computation has received more and more attentions in the field of artificial intelligence; however, the calculation of affective lexicon ontology is a requirement for affective computation of texts. At present, most emotional lexicons are obtained by manual methods. The automatic acquisition of emotional lexicons has become an urgent task that needs to be addressed. This paper presents an automatic acquisition method for emotional lexicons. The authors analyzed the general syntactical rules of emotional lexicons, such as the rules of overlapping words, then rules governing how these acquired emotional words were modified by privatives and degree adverbs. Then we used the conditional random fields （CRF） model to acquire emotional words based on the general rules. Finally, we analyzed the effects of various syntax rules on the automatic acquisition of emotional vocabulary. Experiments were done and the results showed that the proposed method is effective for automatic acquisition of emotional words.
出处
《智能系统学报》
2009年第2期100-106,共7页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金资助项目(60373095
60673039)
国家"863"计划资助项目(2006AA01Z151)
教育部留学人员归国启动基金资助项目(教外留司[2007]1108号)
关键词
情感词汇
词汇自动获取
情感计算
条件随机域
emotional vocabulary
automatic vocabulary acquisition
affective computer
conditional random field
