A hybrid model using data mining and multi-criteria decision-making methods for landslide risk mapping at Golestan Province, Iran

The accurate modeling of landslide risk is essential pre-requisite for the development reliable control and mitigation strategies. However, depends on poorly known environmental socio-economic factors regional patterns occurrence probability vulnerability, which constitute still a matter research. Here, hybrid model described that couples data mining multi-criteria decision-making methods hazard vulnerability mapping presents its application to assessment in Golestan Province, Northeastern Iran. To this end, mapped using three state-of-the-art machine learning (ML) algorithms—Maximum Entropy, Support Vector Machine Genetic Algorithm Rule Set Production—and combine results with Fuzzy Analytical Hierarchy Process computations obtain map. Based obtained results, discussion presented as function main relevant human-environmental conditioning Province. In particular, from response curves algorithms, it can be found p decreases nearly exponentially distance x next road, fault, or river. Specifically, indicated $$p \approx \exp \left( { - \lambda x} \right)$$ where length scale λ about $$0.0797$$ km−1 $$0.108$$ $$0.734$$ 0.734 Furthermore, according follows, approximately, lognormal elevation, while equation = p_{0} K\left( {\theta \theta_{0} } \right)^{2}$$ fits well dependence slope-angle θ, $$p_{0} 0.64,\;\theta_{0} 25.6^{ \circ \;{\text{and}}\;\left| K \right| 6.6 \times 10^{ 4}$$ . highest predicted levels Province are located south southwest areas surrounding Gorgan City, owing combined effect dense local human occupation strongly landslide-prone conditions. Obtained provide insights quantitative risk, priority planning management.