Landslide Prediction with Machine Learning and Time Windows

• Byron Guerrero-Rodriguez ^[1] ; Jose Garcia-Rodriguez ; Jaime Salvador ^[1] ; Christian Mejia Escobar ^[1]
  1. [1] Universidad Central del Ecuador

     Universidad Central del Ecuador

     Quito, Ecuador

     
  2. [2] Universitat d'Alacant

     Universitat d'Alacant

     Alicante, España

     
• Localización: Bio-inspired Systems and Applications: from Robotics to Ambient Intelligence: 9th International Work-Conference on the Interplay Between Natural and Artificial Computation, IWINAC 2022, Puerto de la Cruz, Tenerife, Spain, May 31 – June 3, 2022, Proceedings, Part II / José Manuel Ferrández Vicente (dir. congr.) , José Ramón Álvarez Sánchez (dir. congr.) , Félix de la Paz López (dir. congr.) , Hojjat Adeli (aut.), 2022, ISBN 978-3-031-06527-9, págs. 193-202
• Idioma: inglés
• Texto completo no disponible (Saber más ...)
• Resumen

  • Landslides are among the most destructive natural events, being their prediction necessary to prevent damage to people and infrastructure. This is a problem traditionally addressed with conventional methods, of a deterministic nature, with a limited number of variables and a static treatment of them. In this paper, we propose an approach based on Machine Learning, which has proven to be a successful alternative for dealing with geo-environmental problems. A feature engineering process allowed us to determine the most influential geological, geo-morphological and meteorological factors in the occurrence of landslides. These variables together with the landslide inventory, form a dataset to train different machine learning models, whose evaluation and comparison showed the best performance of the multi-layer perceptron with an accuracy of 99.6%. Our contribution consists of treating precipitation dynamically with the use of time windows for different periods. In addition, we determined the ranges of values of the conditioning factors that combined would trigger a landslide for each time window. Both the multi-temporal prediction and the thresholds of the conditioning factors provide technical support for decision making in risk management.