Robust adaptive genetic K-Means algorithm using greedy selection for clustering

Clustering is a task to divide objects into group depends on their similarity. The optimal of solving clustering problem occurs when the data joins in one group which has a similar category. This study combines Adaptive Genetic Algorithm, K-Means and Greedy Selection to solve clustering problem, named RAGKA. In first step, the centroid is determined by K-Means. Crossover and mutation are performed based on the fitness value of each centroid. At last, the greedy search is operated to get the better solution. To show the performance of RAGKA, five data sets of clustering problem are used. Moreover, RAGKA is compared with other methods as well. The result shows that RAGKA is successfully to solve cluster problem and outperforms than the others.