Text categorization rule extraction based on fuzzy decision tree

In this paper, a new method for text categorization rule extraction based on fuzzy decision tree is presented. An improved chi-square statistic is adopted. The new method reduces features of text in terms of the improved chi-square statistic, and so largely reduces the dimensions of the vector space. And then, a new method for the construction of membership functions is presented, which reduces the time of data fuzzification largely and increase categorization accuracy consequently. Finally, the fuzzy decision tree is applied to the text categorization. Both the understandable categorization rules and the better accuracy of categorization can be acquired.