Study On Reconstruction Of The Enterprise's Competitive Intelligence System Based On Web Mining

With the rapid development of internet, Websites stores a lot of valuable knowledge, building Competitive Intelligence System (short of CIS) has become a strategic measure of lots of enterprises. The main Competitive Intelligence System indexs full text or key words searching mechanism, which brings huge deviation between practical retrieval results and users'demands inevitably. Also the accuracy retrievaled results are hard to control.There are lots of disadvantages of current competitive intelligence extraction methods basing natural language, package learning and ontology-based methods, which is hard to retrieval relevant demands from mass results. Aiming at these problems, we built Web minging Competitive Intelligence System by studying of domestic and foreign retrieval technologies.In this article, we start with retrieving Web page information by using Regular Expressions method to extract original competitive intelligences from mass Web pages flexibly. Considering the existence of half-structure Web pages, which will induce huge redundancy information, the preliminary results should be filtered by data preprocessing technological in order to get rid of advertisement pages and noise datas, then we will get further structural competitive intelligences knowledge.The article includes six chapters, the first chapter depicts current situation of CIS basing on Web mining technology, the second chapter introduces some concepts of CIS basing on Web mining, the third chapter do some demand analysis of CIS basing on Web mining, the fourth chapter describes the key technology of CIS basing on Web mining, the fifth chapter builds a real model of CIS basing on Web mining, and the last chapter realized a implementation of CIS industry.The main achievements as follws: 1) Designed a integrate system of CIS extract ion and minging methods, designed Web page information extraction basing Regular Expressions method, and Web text extraing method based on coding experience regular and CI minging methon based on correlation algorithm. 2) Builded an enterprise Competitive Intelligence System based on Web mining theory and CIS theory. 3) Achieveing a CIS based on the proposed algorithm and model which provides the academic material for further study. In this article, I do the research integrated academic theory and practice; also I have tested the feasibility of the model by implement the CIS what can offer some reference for the enterprises futher research.