The method generates an item set of low dimensionality and a set of classification rules after dealing with the DSM.

• 通过对差异-相似矩阵的处理，最终得到维度较低的文本特征集，并同时生成分类规则。