1:45 PM - 3:15 PM
[SCG55-P15] Reproduction of ETAS model using LSTM
Keywords:LSTM, ETAS, Machine learning
ETAS model (Epidemic-Type Aftershock Sequence model; Ogata, 1988) is one of the models that explain seismicity. Although the ETAS model can explain earthquakes in mainshock-aftershock sequences, it does not explain unusual seismicity such as earthquake swarms.
To solve this problem, we attempted to model seismicity using machine learning. we applied LSTM (Long Short-Term Memory) to earthquake catalogs. LSTM is a type of RNN (Recurrent Neural Network) with the advantage of learning and predicting time-series data. In this study, as a first step to analyzing diverse seismicity, we built an LSTM model that reproduces the ETAS model. For the training data, we used synthetic seismicity data created by the ETAS simulations, which follow the ETAS model perfectly. In the future, we would like to construct a model that can be applied to unusual seismicity that is not explained by the ETAS model.
To solve this problem, we attempted to model seismicity using machine learning. we applied LSTM (Long Short-Term Memory) to earthquake catalogs. LSTM is a type of RNN (Recurrent Neural Network) with the advantage of learning and predicting time-series data. In this study, as a first step to analyzing diverse seismicity, we built an LSTM model that reproduces the ETAS model. For the training data, we used synthetic seismicity data created by the ETAS simulations, which follow the ETAS model perfectly. In the future, we would like to construct a model that can be applied to unusual seismicity that is not explained by the ETAS model.