A hybrid approach for Maximum Entropy clustering using Salp Swarm Algorithm

To address the problem that Maximum Entropy Clustering Algorithm(MEC) is sensitive to the initial clustering centers, we propose a hybrid approach for Maximum Entropy Clustering using Salp Swarm Algorithm (MEC-SSA). First, there are m data samples that are randomly selected as salp populations to get the optimal initial cluster centers. Secondly, the optimal initial clustering centers of MEC are obtained by SSA using DB index as its fitness function. Finally, we can get the better cluster centers by MEC approach. MEC-SSA can alleviate the sensitivity of MEC for initial clustering centers. Further experiments conducted in UCI data show that SSA helps to improve the performance of MEC.