Urban Resilience in the Face of Natural Hazards: Leveraging Machine Learning to Assess Landslide Risk in Kuala Lumpur, Malaysia

Abstract

Landslides represent a significant threat to urban areas globally, causing substantial loss of life, property damage, and infrastructure disruption. The rapid urbanization witnessed in Kuala Lumpur, Malaysia since the 1970s has heightened the susceptibility to landslides, driven by factors such as vegetation removal and slope cutting. This study employs logistic regression (LR), a supervised machine learning technique, to develop a landslide model for Kuala Lumpur. LR is chosen for its simplicity and effectiveness in landslide susceptibility mapping. The methodology involves collecting and pre-processing landslide inventory data, extracting relevant factors from geospatial data, and applying LR to model the relationship between landslides and these factors. The resulting model is validated using an independent landslide dataset, demonstrating a good overall accuracy of 74.1%, with a sensitivity of 84.7% and specificity of 63.5%. The study concludes that LR serves as a valuable tool for landslide hazard assessment and risk management in Kuala Lumpur. The developed model offers guidance for land-use planning and infrastructure development, contributing to Sustainable Development Goal (SDG) 11 by fostering inclusive, safe, resilient, and sustainable cities. By mitigating landslide risk, the model contributes to the protection of lives and livelihoods, promotes sustainable urbanization, and enhances Kuala Lumpur's resilience to natural hazards.