Dynamic Neuro-genetic Weights Connection Strategy for Isolated Spoken Malay Speech Recognition System

Abstract

This paper presents the fusion of artificial intelligence (AI) learning algorithms that combined genetic algorithms (GA) and neural network (NN) methods. These both methods were used to find the optimum weights for the hidden and output layers of feed-forward artificial neural network (ANN) model. Both algorithms are the separate modules and we proposed dynamic connection strategy for combining both algorithms to improve the recognition performance for isolated spoken Malay speech recognition. There are two different GA techniques used in this research, one is standard GA and slightly different technique from standard GA also has been proposed. Thus, from the results, it was observed that the performance of proposed GA algorithm while combined with NN shows better than standard GA and NN models alone. Integrating the GA with feed-forward network can improve mean square error (MSE) performance and with good connection strategy by this two stage training scheme, the recognition rate can be increased up to 99%.