The Research And Implementation Of Data Mining Technology In The Electronic Bulletin Board System Environment

Along with the Internet technology development, the kinds of network application service are more and more. One of them is BBS (Bulletin Boards System), it has provided general network users space to give their opinions freely. But simultaneously some unhealthy or reactionary opinions have brought our society and the country much negative influence. So how to eliminate these Web rubbish accurately and effectively becomes the focus problem which network administrators are concerning over. As the rapidly increasing information, traditional BBS managing method appears backward and ineffectively, which has been not able to adapt the time development. The technology of data mining is just to solve this problem, it can analysis and process large-scale data. Therefore, how to mast this method and use it to realize the BBS safety control effectively is becoming the hot spot that various Websites pay more and more attention.This article, through to describe Web Mining Technique and Text Mining Technique in data mining technology domain, deeply does some researching on the text classification methods. After discussing the text classification method base on vector space model ,we recommend the text classification methode based on class space model. And according to the BBS text feature , we make the improvement on the characteristic extraction algorithm of BBS text classification methods based on the category space model , proposes the combination characteristic extraction algorithm, greatly enhanced the BBS text classification efficiency. Finally we put this new methode into practice, and develope a programe named BBS Content Security Supervision System , simplified the BBS documents data mining processing, and provide a convenient and effective technical method to measure and filter harmful information submitted on BBS. The work we have done include following:1. After analyzing the technology of data mining, web mining and text mining, we put forward a imagination about how to improve the effect on BBS text managing using data-mining technology.2. After comparing the text classification methods based on the vector space and based on the category space, and analyzing the complex of text characteristic extraction, we propose the combination characteristic extraction algorithm in the procedure of BBS text classification based on class space model. The subsequent experiment has proved this method could increase the classification precision and the speed effectively.3. We analyze the BBS constitution, and develope a program named BBS Content Security Supervision System with the mothode of text classification using combination characteristic extraction algorithm Based On Class Space Model.