Machine learning model for the prediction of landslides due to the “El Niño” phenomenon in Peruvian educational institutions

This study develops a machine learning model to predict landslides induced by the “El Niño” phenomenon in educational institutions in Peru. We use a dataset of 55,335 records from the National Center for Estimation, Prevention and Reduction of Disaster Risk (CENEPRED), including geographic and vulnerability characteristics. The model focuses on assessing landslide risk, considering variables such as latitude, longitude, and evacuation plans of institutions, as well as their susceptibility to mass movements. Machine learning algorithms have been used to predict the landslide that may affect the educational infrastructure, obtaining as a result the accuracy of the algorithms Random Forest with 86.23% accuracy, Decision Tree with 83.19%, KNN with 69.85%, MultiLayer Perceptron with 41.49%, concluding that the Random Forest Algorithm has the best accuracy.