The research and application of supervised clustering MOMA

A supervised clustering MOMA (Multi-Objective Memetic Algorithm) was proposed in this paper. In this algorithm, the fitness vector function is the optimization target, and the training samples are gathered according to the similarity of properties into some class clusters supervised by the class label. The number of clusters and the cluster center can be automatically determined by the fitness vector function. The multi-objective genetic algorithm (MOGA) is selected as the global optimization strategy, and the simulated annealing algorithm and the tabu search algorithm are the local search strategies. So those improve the performance of this algorithm. The results of the experiments on the WINE data in UCI validate the feasibility of this algorithm, and the comparison of classification accuracy in credit warning system shows that the prediction accuracy of this algorithm is significantly higher than ANN.