中文摘要：

提出一种微博平台上的产品评论情感倾向性分析方法,对特定语料库的产品特征词的词性进行分析,得到特征词的词性重要程度顺序;以情感词典识别的情感词为起点,提出一种依据词性重要程度的＂特征观点对＂识别规则,并依据该规则进行语义倾向的情感计算。实验结果表明,该方法在产品特征抽取及语义倾向的判断上具有较好的结果。结合微博的转发数和评论数,进行用户观点分析,计算用户对产品的认可度,挖掘产品属性的优缺点,实证分析证明这种方法的可行性。

英文摘要：

The sentiment orientation analysis method in Microblogging platform is introduced.The order of importance of part of speech is derived from analyzing of part of speech of product character word in specific corpus.And starting with recognition of the sentiment word based on sentiment dictionary,a new identification rule for feature-opinion pairs based on the order of part of speech is proposed,and the semantic of sentence is calculated according to the rules.The experiment shows that this method can get a good result in extracting of character of product and computing of semantic orientation of the corresponding of reviews.Combined with forwarding number and comments number of Microblogging,the analysis of users＇ viewpoint in product reviews will be carried out,which calculate the recognition of products,mine the advantages and disadvantages of product attributes.The empirical analysis proves the feasibility of this method.