一种基于决策粗糙集的自动聚类方法

Novel Autonomous Clustering Method Based on Decision-theoretic Rough Set

提出了一种基于决策粗糙集的面向知识的自动聚类方法。在面向知识的聚类算法中,获取初始聚类结果依赖人工阈值的设置。为此,首先根据物理学知识提出了一种差值排序方法来自动得到阈值。另外,讨论了决策粗糙集模型的损失函数,提出了一种聚类评估方法;通过对聚类结果的评估来实现自动聚类。实验结果表明新方法是有效的。

This paper proposed an autonomous knowledge-oriented clustering method based on decision-theoretic rough set model.In order to obtain the initial clustering,the initial threshold values need to set in the knowledge-oriented clustering framework.Thus,a novel method,sort difference,was proposed to produce the initial threshold values autonomously in view of physics theory.Then,a cluster validity index based on the decision-theoretic rough set model was developed by considering various loss functions,which can estimate the quality of clustering.The results of experiments show that the new approach is valuable.