【摘要】 混合属性数据集是现实世界特别是商业金融数据库中最普遍的数据集类型,但适用于这类数据集的聚类算法极少。该文根据聚类融合的方法体系,针对混合属性数据集的特点,提出了基于聚类融合的混合属性特征聚类算法(CEM C),建立了算法框架,列出了算法目标函数和算法主要步骤,并分析了算法复杂度。该算法可以有效处理混合属性海量数据集。用真实数据集验证了算法,并将此算法应用于实际的客户关系管理数据分析中,得到了较好效果。更多还原

【Abstract】 Real-world intelligent databases always have mixed numeric and categorical values which are difficult to cluster.An ensemble-based mixed attribute cluster model was developed for mixed numeric and categorical databases based on the cluster ensemble method.The objective function and the methodology are described in the paper.The method has excellent scalability.Experimental results on real datasets show that the clustering accuracy is better than existing mixed numeric and categorical data clustering algorithms.更多还原