期刊名称：International Journal of Computer Trends and Technology
电子版ISSN：2231-2803
出版年度：2013
卷号：6
期号：2
出版社：Seventh Sense Research Group
摘要：Due to the rapid expansion of digital data , knowledge discovery and data mining have attracted significant amount of attention for turning such data into helpful information and knowledge. Text categorization is continuing to become the most researched NLP problems on account of the everincreasing levels of electronic documents and digital libraries. we present a novel text categorization method that puts together the decision on multiple attributes. Since the most of existing text mining methods adopted termbased approaches, all of these are affected by the difficulties of polysemy and synonymy. Existing pattern discovery technique includes the processes of pattern deploying and pattern evolving, to strengthen the impact of using and updating discovered patterns for looking for relevant and interesting information. But the current association Rules methods exist shortage in two aspects once it is used on patterns classification. a person is the strategy ignored the data about word's frequency in a text . The opposite happens to be the method need pruning rules whenever the mass rules are generated. Within this proposed work specific documents are preprocessed before placing patterns discovery. Preprocessing the document dataset using tokenization, stemming, and probability filtering approaches. Proposed approach gives better decision rules compare to existing approach.
