摘要：Data mining technology has emerged as a means of discovering hidden patterns and trends among large volumes of data and thus it can be considered as an important step in the knowledge discovery (KDD) process. In the area of data mining the task of Association rule (AR) mining is to discover interesting relations among various items in the database. One of the subfield of artificial intelligence is Swarm Optimization which is intended to study the cooperative performance of simple agents. The Particle Swarm Optimization (PSO) is one of the swarm optimization algorithms which can be used for mining improved quality rules. PSO is one of the population based heuristic search technique which is used for solving various NP-complete problems. But PSO has a basic limitation that it gets stuck in local optima. Hence, this research work focuses on; the Binary Particle Swarm Optimization (BPSO) algorithm with cross over operator of Genetic algorithm (GA) for generating better quality association rules among bulky datasets. Due to the better exploration property crossover operator is used with Binary Particle Swarm Optimization (BPSO) algorithm. This algorithm mines improved quality association rules in terms of fitness value without specifying minimum support and minimum confidence thresholds. To prove the practical significance of the approach, this algorithm is tested on three datasets viz. Book dataset, Chess dataset, Connect dataset, using MATLAB and the results obtained has been compared with standard BPSO and GA algorithm. Keywords — Association rule (AR), knowledge discovery (KDD), Particle Swarm Optimization (PSO), Binary Particle Swarm Optimization (BPSO), Genetic algorithm (GA), Support and Confidence.
