中文摘要：

微博情感研究已成为网络文本分析的重要研究领域，微博情感词典是进行微博情感分类的基础。提出一种在分析海量微博语料情感的过程中，自动构建情感词典的方法。方法自动从语料中获取情感词汇、筛选确定情感新词，使用SO-MB 算法计算新情感词的情感极性及强度，构建微博情感词典，结合规则对中文微博进行无监督情感分类。实验证明提出的微博情感词典的构建方法及微博情感分类方法是有效的。

英文摘要：

Research on micro-blog sentiment has become an important research field of network text analysis .Micro-blog sentiment dictionary is the basis for micro-blog sentiment classification .A method is proposed to build micro-blog senti-ment dictionary automatically in the process of analysis massive micro-blog sentiment .The method obtains new sentiment words from the corpus automatically ,then computes the word＆#39;s polarity and the intensity of emotional words by SO-MB and constructs micro-blog sentiment dictionary .Then method uses an unsupervised method to classify the micro-blog sentiment polarity with our dictionary .The experiments show the effective of our method .