Frequent Itemsets Mining Using New Quantum Inspired Elephant Swarm Algorithm

In this paper, an original quantum inspired swarm intelligence approach for solving discrete problems is presented, namely Quantum-inspired Discrete Elephant Herding Optimization (QDEHO). The proposed approach takes advantage of quantum computing characteristics which are studied and analyzed in order to inspire a novel algorithm based on the new Discrete Elephant Herding Optimization (DEHO). As an illustration on how our proposal can be applied on real-life discrete problems, a case study on frequent itemsets mining (FIM) is carried out where the algorithm is modeled and applied on the problem in order to extract interesting patterns from large-scale databases. In order to validate our work, extensive experiments on six relevant datasets with increasing sizes were carried out. The obtained results prove the effectiveness and applicability of our approach. Furthermore, a comparative study of QDEHO with well-known state of the art algorithms such as Particle Swarm Optimization (PSO) and Bat algorithm (BAT) was undertook, where the results showed that QDEHO is superior to the competing algorithms in almost all datasets.