On asymptotic minimaxity of rank tests (Q1200725)

A set of three theorems is developed for obtaining asymptotically minimax lower bounds for testing about functional values. Although the results are developed for rank tests and signed rank tests, they may be used for Kolmogorov, \(\omega^ 2\) and chi-square tests also. The interpretation of rank tests as tests of hypothesis testing about functional values easily extends to the nonparametric case, if we interpret the sets of alternatives and alternative hypotheses as nonparametric.