Diagnosis of Diseases: Classification Rule Discovery from Medical Data using Genetic Algorithm with Suppressor Mutation

Abstract

Data mining is a powerful method to extract knowledge from data. Data mining area focused an attractive research challenges on Knowledge Discovery in Databases (KDD), with the aim to discover interesting and useful data from various transactional databases. Association Rule Mining is one of the most well-known techniques for extracting relations among attributes from larger databases which is given in the form of rules to the user. Data mining techniques in medical domain is inevitable in diagnosing various diseases which leads to reduce time and performance, by taking number of tests required for the patients. This paper presents a rule extraction for diabetes disease using genetic algorithm, with a newly proposed genetic operator named suppressor mutation (GA-SM), to suppress the activity of genes which are over expressive in nature i.e. to suppress the useless attributes from the rules. By using this new operator in association rule mining techniques, it generates a smaller number of rules for diagnosing diabetes. Genetic algorithm shows a wonderful performance with suppressor mutation when compared to CN2, J48, BF tree with respect to performance factors.