Prediction of extreme events in the OFC model on a small world network

TL;DR

This paper demonstrates that in a finite-size small world network modeled by the OFC model, it is possible to probabilistically predict extreme events due to finite size effects, despite the inherent unpredictability in infinite systems.

Contribution

It reveals that finite size effects enable probabilistic prediction of extreme events in the OFC model on small world networks, highlighting the role of system size in predictability.

Findings

Predictability increases with event magnitude and network size.

Finite systems exhibit repulsive temporal correlations of extreme events.

Prediction robustness persists even with limited data.

Abstract

We investigate the predictability of extreme events in a dissipative Olami-Feder-Christensen model on a small world topology. Due to the mechanism of self-organized criticality, it is impossible to predict the magnitude of the next event knowing previous ones, if the system has an infinite size. However, by exploiting the finite size effects, we show that probabilistic predictions of the occurrence of extreme events in the next time step are possible in a finite system. In particular, the finiteness of the system unavoidably leads to repulsive temporal correlations of extreme events. The predictability of those is higher for larger magnitudes and for larger complex network sizes. Finally, we show that our prediction analysis is also robust by remarkably reducing the accessible number of events used to construct the optimal predictor.