Intelligent Modeling and Predicting Investigation of Sound Velocity of Marine Sediments

Physical and mechanical parameters of marine sediment are correlated to sound velocity. At present, there is some deficiency in sound velocity prediction equation with the help of regression fitting method. So it is difficult to obtain sound velocity prediction equation with more than three parameters. This paper applies artificial intelligent method to build a neural network model of multi-parameter sound velocity prediction. The investigation indicates that prediction error of two-parameter-neural-network-model is smaller than the corresponding two-parameter-regression-model, prediction accuracy of which is lower than three-parameter-neural-network-model. This research finding is established as a new way to predict sound velocity of marine sediment.