ON A SERIAL RANK TEST FOR RANDOMNESS AGAINST AUTOCORRELATION
Osamu Sugano · Journal of the Japanese Society of Computational Statistics · 1991
For testing the hypothesis of randomness against the alternative that the variables are autocorrelated, we propose a new serial rank statistic GN and compute the Pitman efficiency of the test based on the statistic GN relative to that based on the autocorrelation coefficient. The serial rank test based on the proposed statistic is applied to some examples.