An enhanced quantum-inspired evolutionary fuzzy clustering

Abstract

Clustering is one of the widely used knowledge discovery techniques to reveal the structures in a dataset that can be extremely useful for the analyst. In fuzzy based clustering algorithms, the procedure acquired for choosing the fuzziness parameter m, the number of clusters C and the initial cluster centroids is extremely important as it has a direct impact on the formation of final clusters. Moreover, the improper selection of these parameters may lead the algorithms to the local optima. In this paper, we proposed an Enhanced Quantum-Inspired Evolutionary Fuzzy C-Means (EQIE-FCM) algorithm to compute the global optimal value of these parameters. In EQIE-FCM, we utilize the quantum computing concept in combination with fuzzy clustering to evolve the different values of these parameters in several generations. However, in each generation these parameters are represented in terms of a quantum bit (Q). At each generation (g), the quantum bit of these parameters is updated using a quantum rotational gate. Through this, after several generations of evolution, we get the global optimal values of these parameters from a large quantum search space. The EQIE-FCM algorithm is applied on the Pima Indians Diabetes dataset and the performance of EQIE-FCM is compared with another Quantum-inspired Fuzzy Clustering (QIE-FCM) and other three fuzzy based evolutionary clustering algorithms from the literature. Extensive experiments indicate that the EQIE-FCM algorithm outperforms many baseline approaches and can be used an effective clustering algorithm. © 2015 IEEE.