Predicting landslide susceptibility based on historical data
Track: Urban and Regional Planning
Authors: Hiran Zani
Landslides are mainly conditioned by environmental variables (i.e., slope inclination, water content, soil and geology) and should be considered for urban planning and management of risk areas. This paper aims to present a method to assess the regional susceptibility of landslides in a portion of the Brazilian territory. Our approach used a multiple logistic regression (MLR) on top of 140 points of historical landslides and produced a map with 86% of accuracy.