The Research Of Association Rule Mining Based On Hybrid Ant Colony Algorithm

Users need to spend a lot of time filtering useful information because of the phenomenon of information overload, which will greatly reduce the user experience. Personalized recommendation technology appears as an effective measure to solve this problem. Currently,association rules mining which is used to solve the problem of individualization recommendation successfully is favored by many researchers. For the problem that recommendation result is less accurate, this paper proposes an association rule mining method based on hybrid ant colony particle swarm algorithm(ACO-MPSO algorithm), so that it can dig out strong association rules that quality are better to improve the accuracy of the results of personalized recommendation.This paper proposes ACO-MPSO algorithm on the basis of association rule mining method based on ant colony algorithm and PSO algorithm, it's main idea is that using the results of PSO algorithm to determine the initial pheromone of colony algorithm to avoid the blindness, and importing metropolis mechanism to avoid the premature phenomenon.According to this idea, this paper designs the implementation process of the ACO-MPSO algorithm in detail, and regards the supermarket data as the data sources, chooses the execution time of algorithm and quality of strong association rules as evaluation indicators,and then compares hybrid ant colony particle swarm algorithm with classical Apriori algorithm, PSO algorithm, ant colony algorithm and hybrid simulated annealing particle swarm algorithm in association rule mining issues. Experimental results show that ACO-MPSO algorithm can dig out strong association rules having better quality.This paper uses association rule mining method based on hybrid ant colony particle swarm algorithm to solve the problem of MovieLens' s personalized recommendation, and regards the hit rate as evaluation index of accuracy rate, and then compares hybrid ant colonyparticle swarm algorithm with Apriori algorithm, hybrid simulated annealing particle swarm algorithm. Experimental results shows that association rule mining method based on hybrid ant colony particle swarm algorithm has higher accuracy.