Adaptive Position–Based Crossover in the Genetic Algorithm for Data Clustering
Arnab Gain, Prasenjit Dey · 2020
This chapter focuses on the optimization of the k-means clustering algorithm. It uses a modified version of genetic algorithm (GA) to optimize the k-means clustering algorithm. The modified version of GA has used a new adaptive position crossover technique to improve the convergence of the GA. The chapter emphasizes the significance of metaheuristic approaches in data clustering. Because of a global optimization technique, the GA has been used in k-means clustering. The performance of clustering is dependent on the initial cluster centers and the outliers present in the dataset. k-means clustering is also stuck in a local minimum. Furthermore, k-means clustering is time-consuming when it is applied to a large dataset. As most of the data clustering algorithms such as k-means clustering suffer from local optima, the GA is needed to obtain global optima. In a few studies, GA-based data clustering has also been used in the case of imbalanced datasets.