Nonparametric Tests of Randomness Against Autocorrelated Normal Alternatives
针对自相关正态备择的随机性非参数检验
Locally most powerful rank tests of randomness against autocorrelated normal alternatives are derived. In particular, autoregressive and moving average alternatives are considered. The distribution of the statistic under the hypothesis of randomness is shown to be asymptotically normal. The small sample power of the test procedure is compared with that of two other procedures by the Monte Carlo Method.