Landslide susceptibility modeling in a complex mountainous region of Sikkim Himalaya using new hybrid data mining approach

Abstract

Landslide is recognized as one of the greatest threats in the complex mountainous regions of Sikkim Himalaya. Therefore, landslide susceptibility modeling (LSMs) has become an ideal tool for managing landslide disasters. Keeping this fact in view, researchers always try to develop optimal models for better performance in LSMs. Thus, the present research study proposed a novel ensemble approach of Alternating Decision Tree (ADTree) and Quantum-Particle Swamp Optimization (QPSO) algorithm and stand-alone of ADTree, QPSO and Random Forest for LSMs in the Rangpo River Basin, India. A total of 342 historical landslide datasets with 14 appropriate landslide causative factors were used for optimal LSMs. The models robustness was appraised via receiver operating characteristics and others statistical indices. Results indicated that QPSO-ADTree model outperformed other models. Overall, the proposed novel ensemble model can be applied as a promising approach for precise LSMs in several complex mountainous regions of the globe.