一种改进的双向聚类算法及其在销售分析中的应用

Modified BiCluster Algorithm and Used in Market Analysis

传统聚类认为每个对象与所有属性均有关系,只是将对象进行聚类,为了克服这一不足,提出基于合并双聚类模型(Merge B iC lusterModel,MBCM)的算法,采用将对象和属性同时进行聚类,找出与一组属性有关的相近对象。实验证明,该方法更为快速,能够得到更加精确的结果。

Traditionary cluster models assume every object has relation to all attributes. Over it and propose a Merge BiCluster Model （ MBCM）, perform simultaneous clustering on the row and column dimensions and get all similar objects assciated some attributes. The experiments demonstrate that the approach is efficient and can get more precise results.