Two-sided Test with Bivariate Normal Response Based on Approximate Likelihood Ratio Method
スポンサーリンク
概要
- 論文の詳細を見る
We consider two-sided tests for testing the superiority and the inferiority between two treatments with bivariate normal response. Although likelihood ratio test has some desirable properties, it is occasionally accompanied by theoretical complications for practical use in multivariate case. Thus we construct a two-sided test by using approximate likelihood ratio method proposed by Tang et al.(1989). We derive formulae for obtaining the critical value and the power of test. Finally we compare our two-sided test with the two-sided test derived by likelihood ratio method proposed by Kudo-Fujisawa (1964) through the simulations.
- 東海大学の論文
著者
-
Douke Hideyuki
Department Of Mathematical Sciences School Of Science Tokai University
-
Imada Tsunehisa
General Education School Of Engineering Kyushu Tokai University
関連論文
- Change Point Analysis for Sequence of Multivariate Normal Means Based on Multiple Comparison Procedure
- Multivariate Multiple Comparison Procedure by Using the Improved Bonferroni Inequality
- A Study on Group Sequential Procedure with Inner Boundaries
- Two-sided Test with Bivariate Normal Response Based on Approximate Likelihood Ratio Method
- Multiple Comparison with a Control for Multivariate One-sided Test Based on the Approximate Likelihood Ratio Test
- Conditional Multivariate One-Sided Test
- Numerical Examples on Multivariate Multiple Comparison with a Control by Using Monte Carlo Simulation
- Multivariate Two-sided Test Based on the Approximate Likelihood Ratio Test
- Multiple Comparison With a Control for Multivariate Normal Means Based on Holm's Method
- Group Sequential Procedure for Bivariate Two-sided Test Based on Approximate Likelihood Ratio Method
- Determination of Different Sample Sizes for Two Treatments in Group Sequential Procedure with Repeated Confidence Boundaries Constructed by Slope
- Group Sequential Procedure with Repeated Confidence Boundaries Constructed by Slope
- Significance Difference Test Between Two Treatments with Changing Population Mean Vectors