Affiliations: Institute of Computer Science and Technology, Chongqing University of Posts and Telecommunications, Chongqing, 400065, P.R. China. yuhong@cqupt.edu.cn | Chongqing R&D Institute of ZTE Corporation, Chongqing, 400060, P.R. China
Note: [] This work was supported in part by the China NNSFC grant(No.61073146) and the Chongqing CSTC grant (No.2009BB2082). Address for correspondence: Inst. of Comp. Sc. and Technology, Chongqing University of Posts and Telecom., Chongqing, 400065, P.R. China
Abstract: In many applications, clusters tend to have vague or imprecise boundaries. It is desirable that clustering techniques should consider such an issue. The decision-theoretic rough set (DTRS) model is a typical probabilistic rough set model, which has the ability to deal with imprecise, uncertain, and vague information. This paper proposes an autonomous clustering method using the decision-theoretic rough set model based on a knowledge-oriented clustering framework. In order to get the initial knowledge-oriented clustering, the threshold values are produced autonomously based on semantics of clustering without human intervention. Furthermore, this paper estimates the risk of a clustering scheme based on the decision-theoretic rough set by considering various loss functions, which can process the different granular overlapping boundary. An autonomous clustering algorithm is proposed, which is not only experimented with the synthetic data and the standard data but also applied in the web search results clustering. The results of experiments show that the proposed method is effective and efficient.
Keywords: clustering, knowledge-oriented clustering, decision-theoretic rough set theory, autonomous, data mining
