Short-time information entropy as a complexity measure

1/f fluctuation observed in many physical, chemical, or biological stochastic processes is expected to be the most efficient for producing "information", or to have the best capability of "computation". In order to measure the complexity of the information included implicitly in the fluctuation, we introduce a unit length of the time series of fluctuation based on a relaxation time of the correlation intensity and define the short-time information entropy of the unit time series. Simulation results show that the short-time information entropy of the 1/f fluctuation is the maximum among f(-v) (0 < v < 2) fluctuations.