机构地区: 衡阳师范学院计算机科学系
出 处: 《计算机工程与科学》 2006年第11期56-59,共4页
摘 要: 本文针对k-modes算法在类的表示方面存在的不足,提出用摘要信息来表示一个类,并给出了一种适用于混合属性的距离定义,得到增强的k-means算法——k-summary算法。理论分析和实验结果表明,k-summary算法较k-modes算法和k-prototypes算法具有更好的精度。 As for the shortcomings of the κ-modes algorithm in the representation of class, we present a method for the representation of class with summary information. A distance definition for mixed attributes is proposed in this paper. Based on the distance definition, we extend the κ-means algorithm and present the κ-summary algorithm, Theoretical analyses and experimental results demonstrate that the κ-summary algorithm can create more accurate class results than the κ- modes algorithm and the κ-prototypes algorithm,
领 域: [自动化与计算机技术] [自动化与计算机技术]