Nonequilibrium Thermodynamics from Time Series Data Analysis

We propose nonequilibrium thermodynamics of non-physical systems such as biological and economical ones1. We inductively construct it based on the Auto Regressive type models2 derived from time series data. Since the detailed balance is generally broken in such non-physical systems, we extend the Sekimoto-Sasa theory3 for systems with the circulation of fluctuations4. We apply our arguments to the time series data of nuclear reactor noise5.