Research Of Association Rules Data Mining Based On Improved Particle Swarm Optimization Algorithm

Association rules mining algorithm was the most core part of the data mining and could find the interesting rules from mass data.Traditional mining algorithm is difficult to adapt to the amount of current data mining.Most of the past researches mainly focused on improving the efficiency of algorithm.However,the confidence and support of the association rules affected the quality of the mining.To improve the efficiency of the mining,this paper presented an association rules mining algorithm based on improved particle swarm optimization(PSO)algorithm.PSO algorithm was easy to fall into local optimal solution when it searched the optimal solution.GSA algorithm could accept bad solution at a certain probability,so it could avoid algorithm trapped in local optimal solution.Combining PSO algorithm with GSA algorithm,an improved hybrid GSA-PSO algorithm was obtained.The new algorithm could not only avoid falling into local optimal solution,but also improve the searching capability of the algorithm.This paper applied the hybrid PSO algorithm to association rules mining and combined the GSA-PSO algorithm with Apriori algorithm for mining association rules,aiming at the problems of the classic Apriori algorithm such as longer running time and easy to extract lengthy rules.It used the real-coded to set the particle.Selecting an appropriate evaluation function applied to the association rules mining and the optimal rules were extracted from the data set.The new algorithm avoided the lengthiness of the apriori algorithm in the extracting rules process and changed longer running time with the increase of data set.By judging the quality of the wine,the simulation analysis proved the feasibility and superiority of the new hybrid algorithm for mining association rules.