The Emotional Tendency Of The Network Of Public Opinion Evaluation Of Technical Research

As our country connected to the Internet, the China's Internet has been rapidly developed. More and more citizens get into the Internet, rapid increase in the number of websites. There are more and more information on the network. As people continue to publish on the network with their points of view, the information on the network with emotion increased rapidly, and the trend is faster and faster.In this paper, the sentiment analysis of the emotional text messages on the Internet is of great significance, the emotional analysis of product reviews will help the company develop the right strategy to win more profits; social media sentiment analysis, will help the government seize the people's need or want and views on policy, conducive to the country's long-term stability.In this paper, I do sentiment analysis on the basis of corpus. This method is not used the traditional classification based on statistics. The traditional approach based on statistical in dealing with emotional issues such as more detailed classification is not good. This paper treat the expression of semantic units as the form of subject-predicate-object which by the method that base on emotional knowledge and linguistics. We make some parse trees and which connected to a multi-loop diagram, with which we do the sentences syntactic process. And then test semantic emotional orientation among formats of every sentence, and ultimately obtained emotional bias of the entire text.The processing units of parse tree is the sentence. In order to improve the accuracy of the parse tree, there also'establish extreme emotional verbs' model, emotion decorated phrase, the subject transfer module, which processing unit is the word level, phrase level, sentence-level context, to fix the result.The paper use a program extracted a candidate set from the corpus according to certain set of rules, then classify the candidate set by manual, and establish the system emotional dictionary.Finally, the portability of system in different corpus has been tested. System which build on the field of Falun Gong, was tested by transfering to the field of hotel rating corpus and obtained good results.