面向中文微博情感分析的多特征融合方法研究

Research on Multi-feature Fusion Method for Sentiment Analysis of Chinese Microbiog

基于中文微博的情感分析一直备受业界关注,但中文微博特征稀疏,语言不规范等特点严重影响了情感分析的质量,基于此,本文提出了一种融合多种特征进行中文微博情感分析的方法,包括基于情感词典的分值特征,基于机器学习的概率特征以及基于深度学习的词向量特征。并通过对照实验验证了该方法的有效性。

sentiment analysis based on Chinese microblogging has been a focus in field of text mining.However, the features of Chinese microblog- ging, such as sparse features and non-standard languages, have seriously affected the quality of sentiment analysis. A sentiment analysis method based on feature fusion is presented in this paper Including dictionary-based sentiment features, machine learning-based Probability features, and Deep learning-based word vector features,Meanwhile, experiments validate them.