Data Clustering Using Chimp Optimization Algorithm

In recent decades, many successful meta-heuristic algorithms are widely applied to data clustering problems due to their powerful capabilities. The Chimp Optimization Algorithm (ChOA) is a recently proposed meta-heuristic search algorithm that is inspired by chimps' individual intelligence and sexual motivation in their group hunting. The good performance of ChOA in a variety of optimization problems, prove the superiority of this algorithm over other swarm-based intelligence. This study presents a novel approach according to ChOA for data clustering. ChOA, like other Swarm Intelligence-based Algorithms (SIAs), starts with a random population and then calculates an objective value for them. In data clustering problems the population of candidate solutions is the position vectors of the centroids, and the objective function is the sum of intra-cluster distances (SICDs) between each sample and its nearest centroid. Following that, ChOA attempts to optimize the population by finding the best position vectors of optimal centroids. There are four different kinds of chimps in this regard: driver, chaser, barrier, and attacker. Furthermore, four major hunting moves including driving, blocking, chasing, and attacking are considered to find the promising solution. In this research, four datasets of the UCI machine learning repository are used to evaluate the performance of the proposed new approach. Experimental results illustrate that the proposed work significantly outperforms other clustering algorithms regarding the objective value, Error Rate (ER), and some other statistical tests.