A symbolic test for testing independence between time series

In this article we introduce a test for independence between two processes {X_t} and {Y_t}. To this end we rely on symbolic dynamics and permutation entropy as a measure of dependence. As a result, a nonparametric (model-free) test for either linear or nonlinear processes is presented. The test is consistent for a broad range of dependent alternatives. Empirical simulations indicate and highlight the general utility of the test for time-series analysts. Copyright Copyright 2010 Blackwell Publishing Ltd