Predicting shear wave velocity of soil using multiple linear regression analysis and artificial neural networks

In this paper, the correlation between shear wave velocity and some of the index parameters of soils including standard penetration test blow counts (SPT), fine-content (FC), soil moisture (W), liquid limit (LL) and depth (D) is investigated. The study attempts to show the application of artificial neural networks and multiple regression analysis in the prediction of the shear wave velocity (VS) value of soils.
New predicting equations are suggested to correlate VS and mentioned parameters based on a dataset collected from Mashhad city in the north east of Iran. The results suggest that better and more exact correlations in the estimation of VS are acquired when ANN method is used. The predicted values using ANN method are checked against the real values of VS to evaluate the performance of this method. The minimum correlation coefficient obtained in ANN method is higher than the maximum correlation coefficient obtained from the MLR. In addition, the value of estimation error in the ANN method is much less than the MLR method indicating the higher confidence coefficient of the ANN in estimating the VS of soil.