On Kolmogorov-Smirnov test


TS Contributor
Let's suppose we have an iid sample of X1,X2,...,Xn.We want to test whether they fit to a distribution f(.|theta) where theta is a vector of the parameters.Plus we want to use the Kolmogorov-Smirnov test.If we estimate theta from the data (using a method like ml,moments etc) does the critical values of the standard test still hold?

I think not. Feedback is soooo needed!