基于语境情感消岐的评论倾向性分析

摘要

研究评论倾向性分析中情感词的动态极性变化问题.用Apriori算法在语境基础上挖掘情感歧义词语搭配,构建出(情感对象,情感词,情感倾向性)三元组形式的情感歧义词搭配词典,利用条件随机场模型(CRFs)序列标注方法从评论文本中抽取出情感要素,在构建的情感歧义词搭配词典基础上对评论文本进行了细粒度情感倾向性分析.在手机和电脑两个领域的评论语料集上进行多组实验,与传统方法的对比实验表明了方法的可行性,较为明显地提高了情感倾向性分析的准确率.%The problem of dynamic polarity change in sentiment analysis was studied.Apriori algorithm was used to expand the sentiment ambiguous words based on context, and constructed the sentiment ambiguous lexicon of triples (namely sentiment object, sentiment word, sentiment polarity).CRFs was used to extracted sentiment elements from comments.Finally, the completed fine-grained sentiment analysis based on the sentiment ambiguous lexicon was conducted.Multiple sets of experiments were performed on two domains of mobile phones and computers.Compared with the traditional method, the experimental results showed the feasibility of the proposed method and the improved accuracy of sentiment analysis.