Generalized P-value Hypothesis Test Of Ratio Of Density And Its Visual And Interactive Window In SAS

The concepts of generalized test variable and generalized p-value were first introduced to deal with one-side hypothesis testing problem. These problems involve nuisance parameters in such a fashion that the derivation of the traditional p-value is not possible. The generalized p-value method has been successfully used to provide solutions for many hypothesis testing problems when nuisance parameters are present. The generalized p-value can also successfully be used to provide small sample solutions for many hypothesis testing problems.In this paper we discusses the one-side hypothesis testing problem concerning ratios of two single-parameter exponential density functions, ratios of two two-parameter exponential density functions, and ratios of two normal density functions. For these one-side hypothesis testing problems, we develop new test procedures with generalized p-value. Secondly, some frequency properties of the generalized p-value are studied by Monte Carlo method which is in the circumstance of SAS statistical software. Thirdly, we present large sample tests based on asymptotic normality of some statistics for the one-side hypothesis testing problem concerning ratios of two single-parameter exponential density functions. Finally, in the result of this paper, a visual and interactive window in SAS statistical software was designed by using the macro and%window programs. The Monte Carlo simulation results of the generalized p-value can be obtained automatically when the necessary parameters for the designed window are filled in the right format.