Increasing Performance of Rule Mining in the Medical Domain Using Natural Intelligence Concepts

This paper discusses how concepts derived from nature can be applied successfully to improve the performance of the rule mining process. These concepts are derived from swarm intelligence and behavior of frogs. Swarm Intelligence (SI) is the property of a system whereby the collective behavior of agents interacting locally with their environment causes coherent functional global patterns to emerge. SI provides a basis with which it is possible to explore collective (or distributed) problem solving without centralized control or the provision of a global model. Association rule mining aims to extract interesting correlations, frequent patterns, associations or casual structures among sets of items in data repositories. Rules have advantages of simplicity, uniformity, transparency, and ease of inference which makes them a suitable approach for representing real world medical knowledge. In this paper, two new algorithms for rule mining have been implemented and their performance has been evaluated over a medical database. Results show that the usability of the rules thus uncovered, is high in the medical domain, and it can be further improved by refining the fitness function. Section I discusses the basic concepts of rule mining and swarm intelligence. Section II describes conventional rule mining techniques and states the motivation behind using swarm intelligence and frog leaping for rule mining and classification. Section III describes the various algorithms that have been implemented in our study. Section IV describes the details of the experiment. Section V presents the results of the practical experiment followed by conclusions and future scope in section VI. KeywordsFitness function, Particle Swarm Optimization, rule mining, rule quality, Shuffled Frog Leaping, Swarm intelligence