Internet Oriented Sentiment Orientation Analysis Of Chinese Public Opinion

With the rapid development of Internet technology and the growing popularity of Internet applications, the Internet has become a very important source from which more and more people obtain information. In the meanwhile, it is becoming a significant platform for people to express their viewpoints. The public opinion on Internet is the sum of beliefs, attitudes, opinions and emotions which most Internet users expressed in variety of social phenomena and problems. Mining and analyzing this rapidly expanding information on web, especially the sentiment of the online reviews posted by users, can better our understanding of the consuming habits and public opinions of various users. Besides, it plays a crucial role in decision-making for many institutions, such as government, the enterprises, etc.At the beginning, this paper introduces the related concepts of public opinions and Internet public opinions and the background of sentiment analysis and its prospect, and takes the pages including reviews crawled by web robot as object, describes the conception and features of Chinese reviews. And then, according to the process of sentiment orientation analysis for reviews, this paper makes a research in the approach of gathering and preprocessing reviews, and the technology of sentiment analysis. For gathering and preprocessing reviews, this paper proposes a detection method based on clustering for review spam. For sentiment analysis, this paper proposes one method based on non-negative factorization for text feature selection.There are a large amount of reviews on public opinions, and there are also some unrelated review spams with the task of sentiment orientation analysis. The review spam interfere the sentiment orientation analysis and affects the accuracy. To solve this problem, this paper designed an unsupervised detection method- a detection method based on clustering for review spam to provide useful review data by detecting and filtering the review spams. Finally, we use the public corpus including Chinese and English as experiment data to verify the validity of this detection method.We always use VSM (Vector Space Model) to express a great deal of texts, and the resulting feature space is too large. It is very necessary to do feature selection to avoid the more use of time and space resources resulted by the too large feature space. We propose one text feature selection method-a text feature selection method based on non-negative factorization after the pre-processing of the texts. We use the existing text classifier to compare the experimentl results based on the new and the traditional feature selection method. The experimental results show that this proposed feature selection method has high accuracy in sentiment orientation analysis of public opinions.