Authors: Changzhong Wang, Yan Wang, Mingwen Shao, Yuhua Qian, Degang Chen

Abstract:

Classical rough set theory is considered as a useful  tool for dealing with the uncertainty of categorical data; the major  deficiency of this model is that the classical rough set model is  sensitive to noise in classification learning due to the stringent  condition of equivalence relation. Thus, a class of fuzzy similarity  relations was introduced to describe the similarity between  samples with categorical attributes. However, these kinds of  similarity relations also have deficiencies when they are used in  fuzzy rough computation. In this paper, we propose a new  fuzzy-rough-set model for categorical data by introducing a  variable parameter to control the similarity of samples. This  model employs the iterative computation strategy to define fuzzy  rough approximations and dependency functions. It is proved that  the proposed rough dependency function is monotonic. Finally, the  proposed model is applied to the attribute reduction of categorical  data. The experimental results indicate that the proposed model is  more effective for categorical data than some existing algorithms.

Keywords： Fuzzy rough set; rough approximation; categorical data; attribute reduction

Fuzzy rough attribute reduction for categorical data.pdf

查看详情

Thu Feb 13 10:50:00 CST 2020