The Research And Implementation Of Sentiment Classification For Tibetan Micro-Blog

Tibetan Micro-blog sentiment classification is designed to mine the sentiment orientation of an event or topic of the Micro-blog users, it is one of the applications of minority language short text analysis research currently. First of all, this thesis analyzes the characteristics of Tibetan Micro-blog; followed by Tibetan Micro-blog sentiment information extraction, construction of Tibetan Micro-blog sentiment corpus; finally through syntactic analysis and a large number of comparative experiments, combined with Tibetan Micro-blog text sentiment characteristics, of Tibetan Micro-blog sentiment classification methods were more in-depth research. The main contents of this paper are as follows:(1) Representation of Tibetan Micro-blog SentimentText representation methods that commonly used are analyzed in this paper. For multi feature vector representation problems existing in the method, put forward a based on semantic space of Tibetan Micro-blog sentiment representation. This method can realize the substitution between the semantic and the feature vector by the syntax tree, and solves the problem that the short text is expressed in a variety of ways and that is difficult to be classified.(2) According to the characteristics of Tibetan Weibo, extraction a variety of labels as a supplement to the emotional information; for Tibetan Weibo in the presence of Tibetan and Chinese mixed arrangement problem, with syntax tree to realize multi language text processing.(3) Build a Tibetan Micro-blog Sentiment Classification SystemTaking Tibetan Micro-blog data basis, Tibetan Micro-blog sentiment classification method based on semantic space as a main classification method, constructed Tibetan Micro-blog emotion classification system. The system can realize Tibetan Sina Micro-blog data acquirement and automatic storage, can also be performed on Tibetan blog sentiment classification, and can achieve high classification efficiency.This paper propose a representation method based on semantic space of Tibetan Weibo emotional, the method through the syntactic tree realized semantic to quantify and improve the emotional characteristics of semantic components, and solves the multi-language text processing is presented in this paper, the problem of the Tibetan Weibo in the presence of Tibetan and Chinese mixed.