文本的情感倾向分析方法

his paper introduces a method targeted at analyzing affective tendency of Chinese texts. First of all, Chinese texts are processed, and then affective words in texts are tagged with the affective words dictionary, and then sentence structure are analyzed. Affective values of the various elements of the sentences have different affective effect to the affective values of sentences; hence all affective elements for the affective values of sentences should be analyzed and weighted. The affective dictionary is based on psychology model, each word has two affective dimensions: activation value, pleasure value. Each word in every dimension has a corresponding value. The affective value of the text is two-dimensional. An affective tendency analyze system targeted at Chinese texts is realized, which consists of a Chinese processing engine and an affective analyzing engine. The affective tendency engine includes affective words identification function, and a rule set of sentences structure. Tests were carried out by using the affective texts.