一种基于近似类抽样的组合聚类方法

FCM聚类算法具有线性的时间复杂度,但它对初始化非常敏感。而k-中心点轮换法对初始化不太敏感,但其缺点就是时间复杂度较高,不能直接应用到海量数据集的聚类分析中。为克服这两类聚类算法的缺点,而充分利用它们的优点,很自然地提出一种基于近似类抽样的组合聚类算法。这种组合聚类算法的时间复杂度是O（n2m）。仿真实验表明,它具有稳定的聚类结果。

An Combination Clustering Method Based on Sampling Using Approximate Aggregation

FCM clustering algorithm tins a linear time complexity, but it is sensitive to initialization. The k- medoids substitution clustering method has better clustering effect and less sensitivity to an initialized medoids set than k - means when clustering those sets of data points with some similarsize clusters. But its time complexity is too high, so it can not be used in huge amounts of data sets. In order to solve their shortcomings, a combination clustering method based on sampling using approximate aggregation is presented naturally. This method needs. This combination clustering method can make a stable clustering effect.