Chaotic analysis of seismic time series and short term forecasting using neural networks

In this study, a chaotic analysis approach was applied to a time series composed of seismic events occurred in Greece. The dynamics of the earthquakes belong to the category of dissipative systems, which exhibit chaotic behavior. After the chaotic analysis, short term forecasting using an artificial neural network has been performed. Neural networks, under appropriate conditions, are known to be universal function approximators, thus they have been used as tools for time series forecasting. Here, a neural network is trained to make short term earthquake predictions. The network architecture is dictated by the calculated characteristics of the time series itself. Preliminary results indicate that this is a promising approach.