A platform for research: civil engineering, architecture and urbanism
A platform for research: civil engineering, architecture and urbanism
Geohazard Sensitivity Evaluation in Xinning, Hunan, China, Using Random Forest, Artificial Neural Network, and Logistic Regression Algorithms
Geohazards may cause great damage to human life and property. It is of great significance to carry out geohazard sensitivity evaluation studies to recognize the development characteristics of geohazards and to predict key hazard areas. In this study, three algorithms, namely, random forest (RF), artificial neural network (ANN), and logistic regression (LR), were used to evaluate the geohazard sensitivity in Xinning, China. Through preliminary remote-sensing interpretation and field validation work, 346 geohazard sites were obtained. Nine influence factors, such as elevation, slope, slope direction, normalized vegetation index (NDVI), rainfall in 2019, distance to rivers, kernel density, distance to faults, and stratigraphic lithology, were selected according to the actual situation of the study area. Three algorithms were evaluated by comparing the size of the area under their receiver’s operating characteristic (ROC) curves (AUC), recall, and F1-score. The results showed that AUC values were 0.8105 (RF), 0.7654 (ANN), and 0.7855 (LR), recall values were 0.7903 (RF), 0.7334 (ANN), and 0.7500 (LR), and F1-score values were 0.7868 (RF), 0.7314 (ANN), and 0.7463 (LR). The RF algorithm had the largest AUC, recall, and F1-score values, indicating that the geohazard sensitivity evaluation model based on the RF algorithm is superior to the ANN and LR algorithms.
Geohazard Sensitivity Evaluation in Xinning, Hunan, China, Using Random Forest, Artificial Neural Network, and Logistic Regression Algorithms
Geohazards may cause great damage to human life and property. It is of great significance to carry out geohazard sensitivity evaluation studies to recognize the development characteristics of geohazards and to predict key hazard areas. In this study, three algorithms, namely, random forest (RF), artificial neural network (ANN), and logistic regression (LR), were used to evaluate the geohazard sensitivity in Xinning, China. Through preliminary remote-sensing interpretation and field validation work, 346 geohazard sites were obtained. Nine influence factors, such as elevation, slope, slope direction, normalized vegetation index (NDVI), rainfall in 2019, distance to rivers, kernel density, distance to faults, and stratigraphic lithology, were selected according to the actual situation of the study area. Three algorithms were evaluated by comparing the size of the area under their receiver’s operating characteristic (ROC) curves (AUC), recall, and F1-score. The results showed that AUC values were 0.8105 (RF), 0.7654 (ANN), and 0.7855 (LR), recall values were 0.7903 (RF), 0.7334 (ANN), and 0.7500 (LR), and F1-score values were 0.7868 (RF), 0.7314 (ANN), and 0.7463 (LR). The RF algorithm had the largest AUC, recall, and F1-score values, indicating that the geohazard sensitivity evaluation model based on the RF algorithm is superior to the ANN and LR algorithms.
Geohazard Sensitivity Evaluation in Xinning, Hunan, China, Using Random Forest, Artificial Neural Network, and Logistic Regression Algorithms
Nat. Hazards Rev.
Wang, Zhengqing (author) / Yu, Yan (author) / Chen, Yulin (author) / Liu, Zhongnan (author) / He, Haiyang (author) / Guo, Zhixin (author) / Ding, Huadong (author) / Li, Ce (author) / Zhang, Dinghao (author) / Peng, Lujun (author)
2025-05-01
Article (Journal)
Electronic Resource
English
Mapping Forest Fire Risk Zones Using Machine Learning Algorithms in Hunan Province, China
| DOAJ | 2023
Forest Fire Probability Mapping in Eastern Serbia: Logistic Regression versus Random Forest Method
| DOAJ | 2020