论文部分内容阅读
提出一种具有特征级别的领域特征集合的情感资源挖掘方法,将基于HowNet词典的分类法构建的情感特征与基于机器学习的特征分类方法中的无内容特征以及领域特征相融合,并将该集合放入支持向量机中进行情感分类实验,实验结果表明,使用抽取模式以及多特征融合的分类方法,可增强中文情感分类效果,验证两种分类方法综合研究的正确性与有效性,弥补目前特征级别的中文情感分类研究的不足。
This paper proposes an emotion resource mining method with feature level set of domain features, which combines the emotion features constructed based on HowNet lexicon and the non-content features and domain features in the machine learning feature classification method, The experimental results show that using the extraction model and the multi-feature fusion classification method can enhance the Chinese affective classification effect, and verify the correctness and validity of the comprehensive study of the two classification methods to make up for the current features Level of Chinese emotion classification research deficiencies.