Discovering High Utility Itemsets Using Swarm Intelligence Algorithm

With the popularity of cloud computing,big data and the Internet of Things,people have ushered in the third wave of informatization.In order to find interesting patterns from these huge amounts of data,data mining methods come into being.Data mining is a bridge to transform data into knowledge,and association rule mining is a specific task of data mining,which aims to discover the possible association or connection between things from data.In association rule mining,frequently common items are called frequent itemsets.the mining problem of high utility itemsets is an extension of frequent item set mining,which not only considers whether the item appears in the transaction,but also uses the weight information of the item,which is more widely applied in real life.Swarm intelligence algorithms take inspiration from nature and model by simulating the behavior of individuals or populations,which has the advantage of not relying too much on specific problems.Different from the traditional data mining methods,the high utility itemsets mining based on swarm intelligence algorithm can be self-organized and self-adaptive iterative updating by combining heuristic rules and random factors to traverse the search space,which shows obvious advantages in solving combinatorial optimization problems.The main research contents of this paper include:1)This paper proposes a high utility itemsets mining method based on Set Particle Swarm algorithm(HUIM-SPSO).Set-Based Particle Swarm Optimization(S-PSO)was used to mine the high utility itemsetst mining.The experimental results verify the advance of the high utility itemsetst mining method based on S-PSO algorithm.In swarm intelligence algorithm,the diversity of population plays an important role in the performance of the algorithm.In the HUIM-SPSO algorithm,in order to measure the diversity of the whole population,the bit-edit distance is proposed,and the maximum bit-edit distance and the average bit-edit distance are defined.Experiments show that the HUIM-SPSO algorithm has higher diversity,and more efficient item sets can be mined within the same number of iterations.2)HUIM-AF,a high utility itemsets mining method based on artificial fish swarm algorithm(AFSA),is proposed.AFSA is a heuristic algorithm with wide application.In the HUIM problem,The artificial fish swarm algorithm only records the current position information without recording additional historical information.The results of the algorithm are not always distributed around several extreme points,which is consistent with the HUIM problem.HUIM was modeled by three behaviors of the artificial fish:following,swarming and preying.The HUIM-AF algorithm is explained in detail,and compared with two related algorithms on four data sets to verify the effectiveness of the algorithm.