摘要
提出一种反映微博文本情感变化的文本情感曲线,采用加窗的方法计算表情符号情感强度,实现自动化的微博表情情感词典构建。其次,基于情感词典和表情词典,计算出反映微博情感变化的微博情感曲线,抽取微博情感曲线波动性、微博情感强度和微博情感倾向性等15种情感特征,采用线性判别分析和贝叶斯分类方法分别对微博进行特征选择和情感分类操作,从而判断微博的情感倾向性。
Proposes a text emotion curve to reflect the emotional changes of microblog text. Shifts the window to calculate emotional intensity of emoticons, develops an automatic construct and update method of emoticon dictionary. Then, based on emotion and emoticon dictionary, calculates the emotional curve to reflect the emotion changes of microblog, extracts the 15 emotion features such as volatility of emotion curve, emotion intensity and emotional bias, uses linear discriminant analysis and Bayesian method to select features and classify emotions, and then judge the emotional tendentiousness of microblog.
出处
《现代计算机》
2015年第14期7-10,33,共5页
Modern Computer
关键词
微博情感分析
情感曲线
表情词典
情感词典
Microblog Emotion Analysis
Emotion Curve
Expression Dictionary
Emotion Dictionary
