Abstract: The huge subjective commenting information on the Internet contains tremendous commercial value and also prospers research on semantic orientation distinction.Sentence semantic orientation distinction is the basic of text semantic orientation distinction.The existing sentence semantic orientation distinction algorithms have some problems，such as the effect of distinction not very well and the difficulty of semantic mode extraction，and so on.This paper looks on the sentiment word as gene.The sentiment word shows different semantic orientations in different contexts.This paper constructs a semantic orientation analyzer of sentiment word.The sentiment word analyzer determines the static semantic orientation of the sentiment word，and then determines the dynamical semantic orientation of the sentiment word.Finally，an algorithm of sentence semantic orientation distinction based on the semantic weighted sentiment word is put forward.The experimental results show the proposed algorithm improves the recall and precision of sentence semantic orientation distinction，and so this algorithm is reasonable and effective.

Key words: sentiment word, context, semantic orientation, orientation distinction

摘要： 互联网上大量的主观评论性信息蕴含着巨大的商业价值，同时也促使了倾向性识别研究的兴起。句子倾向性识别是文本倾向性识别的基础，现有句子倾向性识别方法存在着识别效果不理想、模式抽取困难等问题。将情感词视为基因，在不同的语境下呈现出不同的性状，通过构建情感词语义倾向分析器，先确定情感词的静态显性，然后根据不同的语境确定情感词的动态显性，最后提出基于情感词语义加权的句子倾向性识别算法。实验结果显示，该方法提高了句子倾向性识别的判全率和判准率，是合理和有效的。

关键词: 情感词, 上下文, 语义倾向, 倾向性识别