A Web Text Classification Rules Extraction Algorithm

Text classification is a very important technique for gathering Web information. A novel approach based on multi-population collaborative optimization is proposed for the extraction of Web text classification rules. The information entropy was applied for the initialization of the populations and the multi-population collaborative optimization was applied for the evolution of the populations. The proposed method was applied to three benchmark test sets to examine its effectiveness. Results show that the precision of the proposed method is higher to those of three existing methods, and the cost of computation is less than those of three methods. Furthermore, the classification rules obtained by the proposed method are simple compared with those of three methods.