中文摘要：

用K均值算法进行文本聚类通常只能以局部最优结束，很难找到全局最优．文章提出了一种基于混沌社会演化算法的文本聚类新方法．在该方法中提出了认知主体在聚类中对范式继承的方式，在认知主体对范式的背叛中提出一种混沌变异算子．实验证明该方法不但能有效地提高文本聚类的效率而且能有效地提高文本聚类的精度．

英文摘要：

In the text clustering, K-means clustering algorithm often falls into a local optimum and it is very difficult to find the global optimum. This paper proposes a new text clustering method based on the CSEP （chaotic social evolutionary programming） algorithm. In this method, we present a manner of that a cognitive agent inherits a pamdign in clustering, and a chaotic mutation operator is used for the betrayal of a cognitive agent to a parading. The experiments demonstrate that the present method not only can effectively improve the efficiency of text clustering, but also can effectively improve precision of text clustering.