A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Application of Regression Techniques for Bearing Capacity Prediction in Aizawl
https://orcid.org/http://orcid.org/0000-0002-1893-6689
The bearing capacity of the soil is a critical factor in foundation design. Terzaghi’s method is a widely used method for evaluating the bearing capacity of soil. However, there is a critical need for a model that predicts bearing capacity based on soil properties, particularly for hilly areas where the soil properties vary significantly over short distances due to topography. In this regard, this study aims to develop a data-driven regression techniques-based bearing capacity prediction model using soil index properties as input parameters. Two methods namely stepwise linear regression and Lasso regression were used. Soil samples were collected from 50 different locations in Aizawl City and tested in the laboratory. Thus, obtained geotechnical properties were used for determining the bearing capacity of soil in each location for various types of footing such as strip, square, rectangular and circular footing, respectively. These 50 data points contain bearing capacity as output and geotechnical index properties such as plasticity index (PI), dry unit weight of soil (γ), cohesion (c) and angle of internal friction (ϕ), the width of foundation (B) and depth of the foundation (D) as input and were fed into regression models for training and testing. Performance comparison of stepwise regression with Lasso regression suggested that stepwise regression performed marginally better than Lasso regression in terms of R2 value. RMSE and MAPE values of both the models were comparable. Stepwise performed better than Lasso regression from model complexity point of view with eight (8) predictor terms in comparison to the nine (9) terms in Lasso regression equation. However, based on the input parameters included in the model equation, it may be concluded that Lasso regression has an edge over stepwise regression in simulating the physical process in bearing capacity prediction for our data set. In each site, the square footing has the maximum bearing capacity when compared to the other types of footings. The combination of low PI, higher unit weight of soil and high shear strength parameters can influence high bearing capacity.
Application of Regression Techniques for Bearing Capacity Prediction in Aizawl
https://orcid.org/http://orcid.org/0000-0002-1893-6689
The bearing capacity of the soil is a critical factor in foundation design. Terzaghi’s method is a widely used method for evaluating the bearing capacity of soil. However, there is a critical need for a model that predicts bearing capacity based on soil properties, particularly for hilly areas where the soil properties vary significantly over short distances due to topography. In this regard, this study aims to develop a data-driven regression techniques-based bearing capacity prediction model using soil index properties as input parameters. Two methods namely stepwise linear regression and Lasso regression were used. Soil samples were collected from 50 different locations in Aizawl City and tested in the laboratory. Thus, obtained geotechnical properties were used for determining the bearing capacity of soil in each location for various types of footing such as strip, square, rectangular and circular footing, respectively. These 50 data points contain bearing capacity as output and geotechnical index properties such as plasticity index (PI), dry unit weight of soil (γ), cohesion (c) and angle of internal friction (ϕ), the width of foundation (B) and depth of the foundation (D) as input and were fed into regression models for training and testing. Performance comparison of stepwise regression with Lasso regression suggested that stepwise regression performed marginally better than Lasso regression in terms of R2 value. RMSE and MAPE values of both the models were comparable. Stepwise performed better than Lasso regression from model complexity point of view with eight (8) predictor terms in comparison to the nine (9) terms in Lasso regression equation. However, based on the input parameters included in the model equation, it may be concluded that Lasso regression has an edge over stepwise regression in simulating the physical process in bearing capacity prediction for our data set. In each site, the square footing has the maximum bearing capacity when compared to the other types of footings. The combination of low PI, higher unit weight of soil and high shear strength parameters can influence high bearing capacity.
Application of Regression Techniques for Bearing Capacity Prediction in Aizawl
https://orcid.org/http://orcid.org/0000-0002-1893-6689
The bearing capacity of the soil is a critical factor in foundation design. Terzaghi’s method is a widely used method for evaluating the bearing capacity of soil. However, there is a critical need for a model that predicts bearing capacity based on soil properties, particularly for hilly areas where the soil properties vary significantly over short distances due to topography. In this regard, this study aims to develop a data-driven regression techniques-based bearing capacity prediction model using soil index properties as input parameters. Two methods namely stepwise linear regression and Lasso regression were used. Soil samples were collected from 50 different locations in Aizawl City and tested in the laboratory. Thus, obtained geotechnical properties were used for determining the bearing capacity of soil in each location for various types of footing such as strip, square, rectangular and circular footing, respectively. These 50 data points contain bearing capacity as output and geotechnical index properties such as plasticity index (PI), dry unit weight of soil (γ), cohesion (c) and angle of internal friction (ϕ), the width of foundation (B) and depth of the foundation (D) as input and were fed into regression models for training and testing. Performance comparison of stepwise regression with Lasso regression suggested that stepwise regression performed marginally better than Lasso regression in terms of R2 value. RMSE and MAPE values of both the models were comparable. Stepwise performed better than Lasso regression from model complexity point of view with eight (8) predictor terms in comparison to the nine (9) terms in Lasso regression equation. However, based on the input parameters included in the model equation, it may be concluded that Lasso regression has an edge over stepwise regression in simulating the physical process in bearing capacity prediction for our data set. In each site, the square footing has the maximum bearing capacity when compared to the other types of footings. The combination of low PI, higher unit weight of soil and high shear strength parameters can influence high bearing capacity.
Application of Regression Techniques for Bearing Capacity Prediction in Aizawl
Indian Geotech J
Zirsangzeli, K. (author) / Ramhmachhuani, Rebecca (author) / Mozumder, Ruhul Amin (author)
Indian Geotechnical Journal ; 54 ; 2259-2274
2024-12-01
16 pages
Article (Journal)
Electronic Resource
English
Application of Regression Techniques for Bearing Capacity Prediction in Aizawl
| Springer Verlag | 2024
Corrigendum Notice of NH Circle, Aizawl
British Library Online Contents | 2003
A Semi-Empirical Approach for Rockfall Prediction Along the Lengpui-Aizawl Highway Mizoram, India
| Online Contents | 2022
Planning and Design of Elevated Road for Aizawl City
| British Library Conference Proceedings | 2010