Volume 11, Issue 4 (2-2025)
Journal of Spatial Analysis Environmental Hazards 2025, 11(4): 0-0 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Asghari Sarasekanrood S, sharifi Z, shahbazi Z. Zoning Landslide Hazard in the Masal to Gilvan Road Using a Neural Network Algorithm. Journal of Spatial Analysis Environmental Hazards 2025; 11 (4)
URL: http://jsaeh.khu.ac.ir/article-1-3463-en.html
URL: http://jsaeh.khu.ac.ir/article-1-3463-en.html
1- University of Mohaghegh Ardabili , s.asghari@uma.ac.ir
2- University of Mohaghegh Ardabili
2- University of Mohaghegh Ardabili
Abstract: (212 Views)
Landslides, as one of the most dangerous natural hazards in mountainous regions, continuously threaten human infrastructure, especially roads and transportation routes. Their occurrence often results in significant loss of life and property, making it crucial to study and assess landslide hazards for effective zoning. The purpose of this research is to zone the landslide hazard along the Masal to Gilvan road using a neural network algorithm. The neural network algorithm is recognized as one of the most effective machine learning models, capable of solving complex problems in prediction and classification despite its simplicity. For this zoning analysis, nine influencing factors were considered: (1) geology, (2) vegetation cover, (3) slope, (4) land use, (5) distance from the road, (6) slope aspect, (7) elevation, (8) distance from fault lines, and (9) distance from rivers. The data were prepared, preprocessed, and then entered into MATLAB 2018. A neural network model was designed and implemented with 9 input neurons, 8 hidden neurons, and 1 output neuron. The results indicated that the four most influential factors, ranked by weight, were: slope (0.24), vegetation cover (0.17), distance from fault lines (0.14), and geology (0.11). Final validation using the ROC curve showed that the AUC values were 0.854 for the training phase and 0.971 for the testing phase, both of which reflect highly favorable results. The error rate was found to be very low.
Type of Study: Research |
Subject:
General
Received: 2024/10/13 | Accepted: 2025/03/5 | Published: 2025/03/11
Received: 2024/10/13 | Accepted: 2025/03/5 | Published: 2025/03/11
Send email to the article author
| Rights and permissions | |
 | This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |