Research On Web Text Data Mining

In recent years, with the rapid development of Internet/Intranet, especially the global popularization of web, the present research emphases of data mining transfers from the application on the structure database to the web. Web text mining deals with structure and unstructured data of the web to discover knowledge by use of data mining, machine learning, natural language processing, information retrieval, knowledge management. Text classification is one main research branch of text mining, this paper mainly research the text categorization methodMany techniques have been applied in text categorization, such as the Nearest Neighbore method, Bayesian Networks ,decision trees, neural networks, support vector machines, vector space model, regression model, etc. This paper analyse those common text categorization methods, and then provide a text categorization method based on association rules. First, this paper will introduce the concept and methods of data mining, the skills of Web mining;Second, this paper introduce the mainly algorithms of association rules;Third, this paper introduce the mainly skills used in text data ming , such as the text feature expression method, the text feature extraction method, text classified method;At the end, this paper giving a algorithm of text categorization based on association rules, and proved its feasibility through experiment based on association rules concept.Because both the speed of looking for association rule frequent set and the feature extraction have great effect on classification, so, in the future work, we should promote the researching speed of frequent set, and choose the better feature extraction method to improve our classification.