Volume 1, Issue 2 — Year 2024 — Article e100022

Elmira Ahadi¹; Daryoush Ahadi Rosta²

1. Department of Civil Engineering and Geotechnics, Islamic Azad University, Aras Branch, Hadishahr 5441734335, Iran
2. Department of Civil Engineering, Aeen Kamal Higher Education Institute, Urmia 5715938911, Iran

Landslides are one of the most destructive natural hazards, posing significant threats to infrastructure, human life, and economic activities, particularly in mountainous regions. This study focuses on assessing landslide susceptibility in Azarshahr County, East Azerbaijan Province, Iran, using a combination of Support Vector Machine (SVM) and Logistic Regression (LR) techniques integrated with Geographic Information Systems (GIS). A total of 12 factors influencing landslide occurrence, including slope aspect, distance to the cities, elevation, drainage density, faults’ unsafe radius, rainfall, distance to the river, distance to roads, slope angle, temperature, and weathering, were analyzed. The study employed advanced GIS-based spatial analysis and machine learning methodologies to classify landslide-prone areas into distinct susceptibility zones ranging from very low to very high risk. Results demonstrated the superior predictive capability of the SVM model for handling nonlinear relationships, while LR provided insights into the relative contributions of each factor. The susceptibility map revealed that significant portions of Azarshahr County are categorized as moderate to high-risk zones, particularly in areas with steep slopes and high rainfall. The outcomes offer valuable insights for land-use planning, disaster mitigation strategies, and community safety initiatives. This study highlights the effectiveness of integrating machine learning techniques with GIS for natural hazard assessment, providing a replicable framework for landslide susceptibility mapping in other vulnerable regions.

Landslide susceptibility, Geohazards, Logistic regression, Machine learning, Azershahr