针对客户行为的不确定性和模糊性,将模糊聚类集成技术应用于CRM中的客户细分研究,以提高客户聚类的精度。以模糊C均值(FCM)算法作为基本的聚类器,应用模糊t-范式对生成的多个聚类器进行集成,从而获得最终的客户聚类结果。最后,在10个UCI数据集上进行聚类测试,结果表明,基于模糊t-范式的模糊聚类集成方法的聚类精度要高于常用的客户聚类FCM和K-means方法。在客户信用卡数据集Australian上的学习曲线还表明,聚类集成方法具有更稳定的聚类性能。
Applying fuzzy ensemble cluster to customer segmentation in CRM in order to improve the cluster performance owing to the uncertainty and fuzziness of customer behavior.First,regarding fuzzy C-means algorithm(FCM) as basic clusterer,and then using a fuzzy t-norm to combine the multiple clusterings to obtain the final consensus clustering.Finally,experiments on 10 UCI data sets show that the fuzzy clustering ensemble algorithm based on t-Norm is better than FCM and K-means.Furthermore,experiments on "Australian" data set also indicate that the clustering ensemble algorithm we proposed has better stability.