AutoGluon: A revolutionary framework for landslide hazard analysis

Abstract

Landslide hazard analysis is important to mitigate possible landslide hazards and ensure sustainable development of society and economy. Integrating machine learning models into landslide hazard analysis is a common but challenging task for researchers in general. In this work, we introduce a revolutionary framework, the Amazon's AutoGluon, a new open-source library of machine learning models, to analyze landslide hazards related to the 2017 Jiuzhaigou earthquake in west China. We use 11 mathematical models in the AutoGluon to perform the analysis. For each model, the coseismic landslide inventory and 10 environmental and triggering factors are used as model inputs to perform landslide hazard analysis. These 10 factors are altitude, slope, aspect, slope position, distance parallel to the seismogenic fault, distance vertical to the seismogenic fault, distance to the epicenter, lithology, distance to rivers, and distance to roads. The same number (4,834) of random points in landslide and non-landslide areas are selected, 70 ​% of which are used as training points, and the remaining 30 ​% used as validation points. It takes 47.33 ​s for data preprocessing and model training for 11 machine learning models and the best result measured by Roc-AUC score is 0.94. Our work shows that AutoGluon can greatly improve the efficiency of landslide hazard analysis.