Random Forest based estimate to assess the damages of future earthquakes: preliminary results

Abstract

In this paper we present a case study where the Random Forest (RF) Classifier, has been used to estimate the damage to buildings caused by a (possible) future earthquake, starting from the data of past earthquakes. This prelaminar work is based on the Shakedado dataset, which contains information on buildings and ground shaking parameters for the six major earthquakes that occurred in Italy between 1981 and 2012. We perform the following two conceptual experiments • E1: Assume that the sequence that hit Emilia has just ended and the data relating to the other major earthquakes  happened in the past (L’Aquila, Pollino, and Irpinia) are available, then calculate the level of damage for each building in the Emila dataset. • E2: Assume that the sequence that hit Pollino has just ended and the data relating to the other major earthquakes happened in the past (L’Aquila, Emilia) are available, then calculate the level of damage for each building in the Pollino dataset. Both training and test datasets contain only masonry buildings located within 10 km of the main shock of each sequence. The results show the RF algorithm’s ability to discriminate between buildings with light/no damage from those with medium/severe damage, with a good accuracy, especially for E1.