MVLSC HomeIssue Contents

Mining Approximate Descriptions Using Rough Sets and Genetic Algorithms
Selim Mimaroglu

Using concepts from rough set theory and genetic algorithms we investigate the existence of approximate descriptions of a set of objects, a problem of interest to many fields such as data mining, pattern recognition, machine learning and biology that needs succinct descriptions. A set of objects is defined relative to a given set of attributes; our aim is to find succinct approximate descriptions by removing some of the irrelevant attributes. In this paper, we propose two different methods; first algorithm is based on anti-monotonicity principle and rough sets, and the second algorithm is based on genetic algorithms and rough sets. Proposed methods in this paper have different characteristics and complement each other. Experimental evaluations show that our methods are efficient and practical.

Keywords: Binary Methods, Description Mining, Genetic Algorithms, Rough Sets, Ralaxing Rough Set Border, Fitness Functions

Full Text (IP)