7. Quantum inspired simulated annealing technique for automatic clustering

Clustering is a popular data mining tool whose aim is to divide a data set into a number of groups or clusters. The aim of this work is to develop a quantum inspired algorithm which is capable of finding the optimal number of clusters in an image data set automatically. The article aims to attention on the quantum inspired automatic clustering technique based on a meta-heuristic algorithm named simulated annealing. The quality of this clustering algorithm has been measured by two separate fitness functions named DB index and I index. A comparison has been made between the quantum inspired algorithm with its classical counterparts on the basis of the mean of the fitness, standard deviation, standard error and computational time. Finally, the proof of the superiority of the proposed technique over its classical counterparts has been made by the statistical superiority test named t-test. The proposed technique has been used to cluster four publicly available real life image data sets and four Berkeley image datasets of different dimensions.