ORIGINAL RESEARCH article
Front. Earth Sci.
Sec. Geohazards and Georisks
Volume 13 - 2025 | doi: 10.3389/feart.2025.1619876
Spatial prediction of landslide susceptibility using the data-mining algorithm (Case study: Kamyaran County)
Provisionally accepted- 1Department of Social Sciences, College of Basic Education, University of Halabja, Halabja, Iraq
- 2Department of Geography, College of Human Sciences, University of Halabja, Halabja, Iraq
- 3Department of Geography, Faculty of Education, University of Koya, Koya, Iraq
- 4Department of Forestry, Faculty of Natural Resources, University of Guilan, Someh sara, Iran
- 5Institute of Exploration Technology, Chinese Academy of Geological Science, Chengdu, China
Select one of your emails
You have multiple emails registered with Frontiers:
Notify me on publication
Please enter your email address:
If you already have an account, please login
You don't have a Frontiers account ? You can register here
Mass movements, such as landslides on slopes, are a type of slope activity and a category of natural hazards that result in significant financial, human, and environmental damages globally each year. Identification and classification of regions susceptible to landslides are crucial components of environmental risk evaluation and play a significant role in watershed management. Methods: The aim of this study is to assess the spatial susceptibility of landslides utilizing sophisticated data mining techniques in the Kamyaran County, Iran. Accordingly, the evaluation of landslide susceptibility was carried out employing two advanced data mining approaches, namely Random Forest and Support Vector Machine. In this research, the variables considered for hazard potential zoning included elevation, slope, aspect, slope curvature, distance from rivers, distance from roads, distance from faults, land use, normalized difference vegetation index, lithology, rainfall, and topographic wetness index. A dataset of landslides was utilized for this purpose. The dataset included 103 recorded landslides in Kamyaran County, which served as a map for the actual landslides that took place in the area. To train and validate the models, the landslide data points were split into two subsets, namely training data (70 percent), consisting of 72 points, and validation data (30 percent), comprising 31 points. Ultimately, the efficacy of the models was assessed using the receiver operating characteristic (ROC) curve.The findings from the ROC curve analysis revealed that the SVM and RF models achieved AUC values of 0.91 and 0.95, respectively; thus, the RF model exhibited the highest AUC value in comparison to the SVM, making it the most effective model for forecasting landslide susceptibility in the study area in the future.Landslide potential maps are valuable tools that can be applied in environmental management, land use planning, and infrastructure development.
Keywords: landslide, Data Mining, random forest, Support vector machine, Spatial prediction
Received: 28 Apr 2025; Accepted: 29 May 2025.
Copyright: © 2025 Ashraf Mohammed, Kamal Hamaamin, Azad, Rasooli and Li. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
* Correspondence: Houzhi Li, Institute of Exploration Technology, Chinese Academy of Geological Science, Chengdu, China
Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.