|
Enhanced prediction of slope stability and failure distance using hyperparameter tuning and polynomial features |
|---|---|
| รหัสดีโอไอ | |
| Creator | Pattanasak Chaipanna |
| Title | Enhanced prediction of slope stability and failure distance using hyperparameter tuning and polynomial features |
| Contributor | Pornkasem Jongpradist, Jirawat Supakosol, Piyoros Tasenhod, Raksiri Sukkarak, Nattawut Hemathulin |
| Publisher | Mahasarakham University |
| Publication Year | 2569 |
| Journal Title | Journal of Science and Technology Mahasarakham University |
| Journal Vol. | 45 |
| Journal No. | 1 |
| Page no. | 109-121 |
| Keyword | Slope stability analysis, machine learning, slope failure distance |
| URL Website | https://li01.tci-thaijo.org/index.php/scimsujournal |
| Website title | Journal of Science and Technology Mahasarakham University |
| ISSN | 1686-9664 (Print), 2586-9795(Online) |
| Abstract | This study presents an extensive analysis of slope stability using various machine learning (ML) models, focusing onhyperparameter tuning and feature importance and model validation for predicting factor safety (FS), slope failuredistance relative to the height of the slope, and the safety level of the slope. The input parameters include cohesion(C), internal friction angle (Phi), slope angle (Slope), and height of the slope (H). The performance of each model wasassessed using Mean Absolute Percentage Error (MAPE), Mean Squared Error (MSE), and R-squared (Rฒ) forregression tasks, while classification tasks were evaluated using accuracy, precision, recall, F1-score, and Area Underthe Curve (AUC). The analysis demonstrates the efficacy of different ML models. The model that performed well in both regression and classification is the Random Forest. The use of polynomial features significantly improved theperformance of linear methods, while hyperparameter tuning greatly enhanced the performance of Support Vector and MLP models. |
