Non-parametric test for colored noise in time series data (MATLAB)

This code is an implementation of the method by Kennel and Isabelle (Physical Review A, 1982) to test the null hypothesis of the time series having ths same generating distribution as a colored noise process. The method computes a Kolmogrov-Smirnov test statistic from a deterministic, time-delay embedded time series by comparing the prediction errors of surrogate (noise added) and original time series data. Under the null hypothesis the test statistic should have standard mean and unit variance.

Wikimedia Commons

Download code here