| Subjective Chinese text analysis and segmentation has a lot of research and application value, sentiment searching aroused the interest of many researchers in recent years. Sentiment searching is a process which mining and analyzing the subjective information generated by users on the Internet. Studies of sentiment searching mainly focus on the process of subjective text analysis, also known as sentiment analysis or opinion mining. By analyzing and mining subjective information, sentiment analysis obtained useful subjective knowledge and information. At present, most sentiment analysis technologies are based on word-level analysis, however, the performances of Chinese words segmentation method will affect the result of sentiment analysis. Therefore, research the affection of words segmentation method for sentiment analysis is very important.This paper firstly summarizes the development and research status of Chinese sentiment analysis and Chinese words segmentation, then combine the technical characteristics of Chinese words segmentation and sentiment analysis technology, use Rule-based segmentation methods and statistical-based segmentation methods, propose Sentiment Segmentation Method, established the Sentiment Segmentation Model. The experimental results show that the segmentation method enhanced the segmentation precision of sentiment words. Secondly, according to the procedure of sentiment classification, this paper compare four kinds of feature extraction methods, three kinds of weight calculation methods and two kinds of classification methods; Finally, according to sentiment classification methods, this paper establish a new sentiment classification model, validating and testing the performances of Sentiment Segmentation Method. This paper interspersed describes the evaluation and related resources of Chinese text sentiment analysis. Finally, we summarize the research difficulties of sentiment searching, and the future direction of sentiment searching. |
PDF Full Text Request