[U-04] Geospatial Applications for Societal Benefits

[U04-10] Machine Learning for Urban Resilience: Landslide Susceptibility Assessment in Klang Valley, Kuala Lumpur

*Mohammed Arkan Majeed¹, Elanni Md Affandi¹, Chin Yik Lin¹, Noer El Hidayah Ismail¹, Mohd Talha Anees¹, Harry Telajan Linang¹ (1.Universiti Malaya)

Keywords:landslide susceptibility, machine learning, tropical climate, urban resilience, LiDAR

Klang Valley, Kuala Lumpur, Malaysia, is increasingly facing landslide risks to public safety and infrastructure caused by factors such as heavy rainfall, steep terrain, and rapid urbanization. Therefore, to reduce landslide risks, effective landslide susceptibility assessment is of paramount importance. Although machine learning models and lidar technologies are promising, they require further investigation for their optimal use, especially in the unique tropical environment and climate. Leveraging historical landslide inventory and building on bivariate statistical methods that achieved 92% accuracy in a previous study conducted in Kuala Lumpur, this research compares the effectiveness of machine learning models in terms of model performance, selection, and weighting of influential parameters. In addition, this research explores the capabilities of machine learning to predict high-risk landslide areas based on historical data. This study will establish a comparative understanding of the methods and data used in machine learning-enabled landslide susceptibility mapping and disaster management in tropical urban environment.