chee chen
2014-Mar-07 18:57 UTC
[R] Question: How Kolmogorov Test is computed by R function "ks.test", using scaled difference or unscaled
Dear All, I would like to ask for your help on the following when testing uniformity by the KS test: 1. We know that the Kolmogorov test is based on the supremum D of the difference between the empirical distribution and the uniform distribution 2. We also know that sqrt(n)D has a limiting distribution K, where n is the sample size 2. However, when we use the R function "ks.test" to compute the p-value of two-sided KS test, is D fed into K or sqrt(n)D fed into K? Thank you for your time and effects into my question. Che [[alternative HTML version deleted]]