一种基于粗糙集的K-Means聚类算法

A Rough-based K-Means Clustering Algorithm

在传统的硬聚类过程中,得到的簇中数据对象是确定的,然而在现实世界,边界数据是不能被准确划分到任何一个簇的。粗糙集是处理这种边界不确定性的工具,基于此提出了一种基于粗糙集的K-Means聚类算法,这种算法生成的簇包括上近似集和下近似集,可以处理边界对象。试验证明,这种算法是有效的。

Traditional hard clustering get crisp boundary of each cluster,however,in reality the boundary of data is hard to partition.Rough set is an effective tool to deal with the uncertain boundary region.Based on it,the paper puts forward a rough-based K-means algorithm which gets the clusters including lower approximations subsets and upper approximations subsets to deal with boundary region.Data experiment shows the effectiveness of the algorithm.