In this paper we present the preliminary results of a study using a deep-learning tool named LSTM (Long Short-Term Memory) network, to classify seismic events as near-source and far-source, with the final purpose of developing efficient earthquake early warning systems. We use a similar approach as in [15], applied to a database, named Instance, containing information about 54,008 earthquakes that occurred in Italy. Although these are preliminary results, the method shows a good ability to detect far-source events with an accuracy of about 67 %. For near-source events, the method shows an improvable result with an accuracy of 57 %. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.
A Preliminary Result of Implementing a Deep Learning-Based Earthquake Early Warning System in Italy
Adebowale A.;Di Michele F.;Rubino B.
2023-01-01
Abstract
In this paper we present the preliminary results of a study using a deep-learning tool named LSTM (Long Short-Term Memory) network, to classify seismic events as near-source and far-source, with the final purpose of developing efficient earthquake early warning systems. We use a similar approach as in [15], applied to a database, named Instance, containing information about 54,008 earthquakes that occurred in Italy. Although these are preliminary results, the method shows a good ability to detect far-source events with an accuracy of about 67 %. For near-source events, the method shows an improvable result with an accuracy of 57 %. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.Pubblicazioni consigliate
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
