Association rules optimization using improved PSO algorithm

In this work, association rules are optimized by using improved particle swarm optimization algorithm (PSO Algorithm). Here improved PSO algorithm means classical PSO algorithm with additional operator in the forms of mutation of genetic algorithm. The basic shortcoming of PSO algorithm is to get trapped into local optima. So for improving this, mutation operator is used additionally in classical PSO algorithm. This operator is used after the initialization phase of PSO algorithm. Firstly, different association rules for generating frequent item sets are generated by standard Apriori algorithm, then improved PSO algorithm is applied on these generated association rules for optimizing them. Experiments are performed on different datasets taken from UCI machine learning repository and results are compared with other previously proposed algorithms, called KNN algorithm and ABC algorithm. These results show that the proposed algorithms efficiency is better than previously proposed algorithms.