Fast Reduction Method in Dominance-Based Information Systems
Abstract
In real world applications, there are often some data with continuous values or preference-ordered values. Rough sets based on dominance relations can effectively deal with these kinds of data. Attribute reduction can be done in the framework of dominance-relation based approach to better extract decision rules. However, the computational cost of the dominance classes greatly affects the efficiency of attribute reduction and rule extraction. This paper presents an efficient method of computing dominance classes, and further compares it with traditional method with increasing attributes and samples. Experiments on UCI data sets show that the proposed algorithm obviously improves the efficiency of the traditional method, especially for large-scale data.
- Publication:
-
Materials Science and Engineering Conference Series
- Pub Date:
- January 2018
- DOI:
- 10.1088/1757-899X/301/1/012134
- Bibcode:
- 2018MS&E..301a2134L