Effective utilization of optimization techniques in artificial neural networks for speaker recognition