CODA WAVE ANALYSIS USING ARTIFICIAL NEURAL NETWORKS

Abstract

In this dissertation we propose the use of artificial neural networks in the analysis of coda waves. This approach significantly reduces the time and work in the analysis of coda waves compared to other methods of coda wave analysis. The comparison of various neural network learning algorithms as well as different activation functions is done. The layer recurrent neural network with Purelin activation function gives the minimum error in the testing phase and thus pro'es to be the best among the neural networks used to analyze the coda waves.