Statistical Modeling to Predict Phase Velocity Parameters Based on Active Multichannel Analysis of Surface Waves Soil Profiles at Geotechnical Sites
A. Arisona *
Geological Engineering Department, Faculty of Earth Science and Technology, Halu Oleo University, Kendari, Indonesia.
K. S. Ishola
Department of Geosciences, University of Lagos, Akoka, Lagos, Nigeria.
. Muliddin
Geological Engineering Department, Faculty of Earth Science and Technology, Halu Oleo University, Kendari, Indonesia.
. Hasria
Geological Engineering Department, Faculty of Earth Science and Technology, Halu Oleo University, Kendari, Indonesia.
L. D. Ngkoimani
Geological Engineering Department, Faculty of Earth Science and Technology, Halu Oleo University, Kendari, Indonesia.
. Bahdad
Geological Engineering Department, Faculty of Earth Science and Technology, Halu Oleo University, Kendari, Indonesia.
L. D. Restele
Department of Geography, Haluoleo University, Kendari, Indonesia.
*Author to whom correspondence should be addressed.
Abstract
The inversion of MASW dispersion curves were performed using Dinver-Geopsypack to obtain phase velocity. The profile was selected for the MASW analysis. The objective of statistical modeling in this study is to verify the field measurement model of the phase velocity parameters using a validation procedure. Model validation has been made to understand if the model created represents the field observation. These involve training and testing procedures of the models to identify the existence of soil profiles for geotechnical site investigation. The procedure is carried out to select the MASW parameter estimation model in terms of phase velocity to determine which one best fits the data with high determinant coefficient values, parameter estimates and low root mean square error (RMSE) values. Furthermore, analysis of variance (ANOVA) known F-test and significance – α were carried out to establish a confidence level of 95% and 99%. The calculated F-values are greater than the tabulated F-values, suggesting that the developed models are statistically valid. The results of the statistical modeling showed a positive significance (best-fit) was achieved between the prediction and the measured data (observations), and confirmed the accuracy of the model, especially the phase velocity for predicting the soil profile.
Keywords: ANOVA, RMSE, MASW, phase velocity, significance