A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Prediction of Spatial Soil-California Bearing Ratio of Subgrade Soil Using Particle Swarm Optimization—Artificial Intelligence Method
This research explores the integration of elevation above sea level in California Bearing Ratio (CBR)-based road design to improve the accuracy of subgrade condition assessment. CBR is considered crucial for flexible pavement design but is often hindered by the limitations of traditional statistical models. This study employed five hybrid artificial intelligence (AI) models: Particle Swarm Optimization (PSO) with KNearestNeighbor (KNN), GradientBoostingRegressor (GBR), RandomForestRegressor (RFR), SupportVectorRegressor (SVR), and RidgeRegression (RR). Using a dataset of 509 data points from ongoing road projects. The elevation, moisture content, and bulk density parameters were used as input for the first time together. The PSO-GBR hybrid model emerged as the most accurate for testing, achieving a coefficient of determination (R2) = 0.97, root mean squared error (RMSE) = 0.7, mean absolute error (MAE) = 0.46, and mean squared error (MSE) = 0.5. The performance of built-in models is measured by three new index performance metrics of GB-R: the a20-index = 99.02, the index of scatter (IOS) = 0.13, and the index of agreement (IOA) = 0.99, in addition to four common metrics. Feature selection and feature importance analysis further confirmed swelling and elevation’s critical role in improving predictions. Offers better insight into the influence of environmental factors on infrastructure design and soil behavior.
Prediction of Spatial Soil-California Bearing Ratio of Subgrade Soil Using Particle Swarm Optimization—Artificial Intelligence Method
This research explores the integration of elevation above sea level in California Bearing Ratio (CBR)-based road design to improve the accuracy of subgrade condition assessment. CBR is considered crucial for flexible pavement design but is often hindered by the limitations of traditional statistical models. This study employed five hybrid artificial intelligence (AI) models: Particle Swarm Optimization (PSO) with KNearestNeighbor (KNN), GradientBoostingRegressor (GBR), RandomForestRegressor (RFR), SupportVectorRegressor (SVR), and RidgeRegression (RR). Using a dataset of 509 data points from ongoing road projects. The elevation, moisture content, and bulk density parameters were used as input for the first time together. The PSO-GBR hybrid model emerged as the most accurate for testing, achieving a coefficient of determination (R2) = 0.97, root mean squared error (RMSE) = 0.7, mean absolute error (MAE) = 0.46, and mean squared error (MSE) = 0.5. The performance of built-in models is measured by three new index performance metrics of GB-R: the a20-index = 99.02, the index of scatter (IOS) = 0.13, and the index of agreement (IOA) = 0.99, in addition to four common metrics. Feature selection and feature importance analysis further confirmed swelling and elevation’s critical role in improving predictions. Offers better insight into the influence of environmental factors on infrastructure design and soil behavior.
Prediction of Spatial Soil-California Bearing Ratio of Subgrade Soil Using Particle Swarm Optimization—Artificial Intelligence Method
Transp. Infrastruct. Geotech.
Tilahun, Yonas (author) / Xiao, Qinghua (author) / Ashango, Argaw Asha (author) / Han, Xiangyu (author) / Negewo, Mesfin (author)
2025-01-01
Article (Journal)
Electronic Resource
English
Correlation of California Bearing Ratio (CBR) Value with Soil Properties of Road Subgrade Soil
| Online Contents | 2018
Correlation of California Bearing Ratio (CBR) Value with Soil Properties of Road Subgrade Soil
| Online Contents | 2018
Prediction of California Bearing Ratio from Soil Index Properties Using Artificial Neural Network
| Springer Verlag | 2024
Onsite Estimation of California Bearing Ratio of Subgrade Using Sensor Acceleration
| Springer Verlag | 2024