Testing for general dynamical stationarity with a symbolic data compression technique

We construct a hypothesis test for examining the stationarity of the evolut ion law for a time series of discrete symbols: whether two data streams app ear to originate from the same underlying, but unknown, dynamical system. B ased on techniques from the theory of data compression, our method intellig ently accounts for the substantial serial correlation and nonlinearity foun d in realistic dynamical data. We demonstrate the method on a number of rea listic experimental datasets.