Analysis And Research Of Text Orientation Based On Semantic

In recent years, with the rapid development of Internet technology, more and more network information emerged. However, it is not just that the Internet is access to information repositories, but also to express their views and exchange views on the garden. For Internet users, they can publish comments and suggestions for a hot event, or understanding the quality and services of products to relevant enterprises through the Internet; For enterprises, they can create product reviews column, conducting market research and analysis, understanding the views and comments of users to the products, so as to improve product performance and after-sales service.In order to extract the useful information from rich information, The analysis importance of text orientation have been highly valued. The main purpose is to judge the emotional orientation which critics pose to object. The main work of this paper is summarized as follows:(1) Describe the methods of the traditional text orientation analysis and point out the deficiency. Through the analysis of semantic information and the theoretical knowledge of semantic orientation, we discuss three kinds of semantic analysis methods based on semantic orientation.(2) Based on the basic HNC theory, the method semantic-correlation which is based on the system of HNC concepts primitive symbols is presented, then proposed the calculation strategies for semantic-correlation. Finally, the semantic-based HNC correlation method is applied to calculate the words of the original polarity.(3) Propose a text orientation algorithm which based on I-SBV polar transfer. first, description of the overall framework of the text orientation algorithm, then, analysis of three current text orientation algorithm and improvement of the SBV polar transfer algorithm due to the characteristic of the network comments. Last , proved of the effectiveness of the improved algorithm.(4) Designed the whole restaurant reviews system which based on I-SBV transfer algorithm, and introduced the main modules in detail. Analysis of the four characteristics (price, dishes, environment, services) on the restaurant in the paper, and definition of several indicators to measure the results of evaluation.