One-side Hypothesis Testing Problem Based On Empirical Likelihood

In this paper, we consider a one-side hypothesis testing problem H₁₀ :μ=μ₀ vs H₁₁:μ>μ₀; H₂₀:μ≥μ₀ vs H₂₁ :μ<μ₀ for the population meanμby the method of empirical likelihood, and present the asymptotic distribution of the statistic of nonparametric likelihood ratio under the null hypothesis,respectively. Our simulations show that the powers of the statistic of nonparametric likelihood ratio are much lager than those of Wilcoxon's rank statistic.