n-Mating Effect on Genetic Algorithm-Based Clustering Performance for Hotspots Data

This study aims to explain the effect of variations in the number of individuals mated with father (n-mating) on the performance of Genetic Algorithm Polygamy (GAP) clustering. GAP clustering is clustering method based genetic algorithm. The steps of this method are same as GA clustering steps, but the crossover process is done with polygamy. One selected father is mated with more than one mother. We evaluate the performance of GA-based clustering for hotspot data with three clustering evaluations, namely Sum Square Error, Davies-Bouldin Index, and Silhouette Coefficient. Based on experimental result, GA Polygamy clustering outperforms GA clustering based on Sum Square Error (SSE) evaluation and Silhouette Coefficient (SC) evaluation. The n-mating in the crossover process of GAP clustering affects GAP clustering performance also performance of time to convergence of GAP clustering.