The Research And Application Of Rough Set Based On Information Entropy

With the increasingly large Internet information resources, the rapid acceleration of information transmission speed, Internet services to provide people with more convenient way, to constantly enrich the content.For example, people can publish on the Web blog, can share all the interesting things they know with all the blog friends, blog friends can also share other things to comment, to express their views and opinions.People can transactions on the network, buying and selling, and even evaluate the quality and price of traded products, the integrity of buyers and sellers.Because of the virtual characteristic of the network activity, whether we are on the network to share anecdotes or trading activities, people can not see the existence of things or the things that really, only be able to see things on the web descriptive existence. In order to further deepen understanding of such network existence things, people are more inclined to refer to the network through the existing network evaluation of such things to multi-faceted, multi-perspective knowledge and understanding of such network exists in all aspects, such as the authenticity of the anecdotes, online trading product quality, cost-effective, the integrity of transaction double degree which introduced before.However, due to network review eyes of the beholder wise see wisdom, and as to the same thing, people may form different comments, but the results there are two, positive or negative. How can a computer system to analyze comments on these networks to determine the tendency of people to evaluate things, that the Chinese text of these feelings to determine the tendency of Internet users is undoubtedly of great value.Orientation of the text is the scope of computational linguistics. In the computational linguistics and related field,the study on the objectivity of the information is relatively small, and the subjective analysis and extraction of information is not a lot, still in its infancy, there are many problems to be a comprehensive exploration.The study involves artificial intelligence, machine learning, information retrieval,data mining, and many other basic research. Rough set theory is a theory proposed by a Fellow of the Polish Academy of Sciences, Z. Pawlak in 1982, which is used in data analysis and reasoning.The main task of rough set theory is to approximate classification, knowledge Reduction (Reduction attributes and attribute values), attribute dependency analysis, generated under optimal and sub-optimal decision-making control algorithm decision table basis on the decision table.In this paper, we pretreatment the web reviews throught the analysis of the form characteristics of the network comment statements,combined with attribute reduction of rough set theory,information entropy, pattern matching algorithms,and proposed a kind of rough set theory which is based on information entropy to judge the tendency of the Chinese text,and finally we Verified the feasibility of this method by experiment.