원문정보
초록
영어
There are fewer techniques to group objects having similar characteristics deal with categorical data ,but some are of them be complicated in the clustering process while others have stability issues. In this paper we represent a new technique which it be more easier than the other techniques in computing the selecting clustering attribute process and at the same time having stability issues besides taking care of handling uncertainty and categorical data together, we called it (maximum significance of attributes) MSA. The proposed technique based on rough set theory by taking into account the concept of significance of attributes of the database. We analyzing and comparing the performance of MSA technique with (bi-clustering) BC, (total roughness) TR, (minimum-minimum roughness) MMR and (maximum dependency of attribute) MDA techniques.
목차
1. Introduction
2. The Main Concepts of Important Definitions
3. Proposed Algorithm
4. Experimental Part
4.1. Computational Part
5. The performance comparisons of MSA with that of BC, TR, MMR and MDA techniques
5.1. Objects splitting for TR, MMR and MDA techniques
5.2. The Purity Ratio for TR, MMR and MDA Techniques
5.3. Objects Splitting for MSA Technique
5.4. The Purity Ratio for MSA Technique
6. Conclusions
Acknowledgement
References