Firefly Algorithm For Optimization of Association Rules

With the remarkable increase in online transactions, market basket analysis has become more relevant. Identifying the buying patterns and generating more relevant recommendations to users is the prime goal of online shopping portals. However, extracting associations between products from millions of transactions is a big challenge. Association rule mining techniques are commonly used to discover the associations among transactional databases. To further enhance the elicitation of important strong rules between items, this work applies metaheuristic algorithms such as Genetic Algorithm, Particle Swarm Optimization and Firefly algorithm on Apriori algorithm. The performance of proposed approach is evaluated in terms of number of rules generated and the time complexity of the algorithm. Experiments performed over two different transactional datasets of varying size reveals that Apriori with firefly results into least number of rules as compared to GA and PSO.