Analysis of feed forward and recurrent neural networks in predicting the significant wave height at the moored buoys in Bay of Bengal

In meteorological and marine engineering applications prediction of Significant Wave Height (SWH) plays a major role for forecasting cyclones, earthquakes & tsunamis that may occur in the ocean and warn the society for appropriate action. Recently, researchers are exploring the use of soft computing techniques to predict SWH In this work, the wind and wave data obtained from moored buoys of Bay of Bengal is used to predict the SWH using Artificial Neural Networks (ANN). The Recurrent and Feed Forward Networks were analyzed using Levenberg Marquardt (LM), Conjugate Gradient (CG) and Bayesian Regularization (BR) algorithms. Results indicate that Recurrent Networks with BR algorithm has higher correlation and less error.