Study of Rough Set Text Categorization Techniques Based on Combined Extracting Features
Kejun Zhang · Jisuanji yingyong yanjiu · 2007
The focus of the research was to extract text features.Through mutual information and χ2 function study in accor-dance with their respective errors,introduced a new feature extraction algorithms: CEFA.Through this algorithms could extract more representative characters.Using the superior reduction of the rough sets to construction a text categorization system,to extract decisionmaking rules for the text categorization.The experiments show that the categorization accuracy is higher.