期刊名称:International Journal of Computer Science and Information Technologies
电子版ISSN:0975-9646
出版年度:2012
卷号:3
期号:2
页码:3733-3736
出版社:TechScience Publications
摘要:Rule generation in an expert system requires heuristics and selection procedures which are not just accurate but are also efficient. This premise makes Genetic Algorithms (GAs) a natural contender for the rule selection process. The work analysis the previous attempts of applying GAs to rule selection and proposes major changes in the clustering algorithms for rule generation. The representations of a cluster, the formula for probability and the term distance have been modified. Also the coefficient of auto correlation and not mean is taken as the deciding factor for an item to be in the cluster. The work has been implemented and analyzed and the results obtained are encouraging.