Test Statistic Distribution of Composite Hypothesis with Parameter Space Restriction in Continuous Multivariate Distribution

The null hypothesis testing of linear combination of p-dimensional parameter vector associated with an known and full rank matrix against the one sided linear combination of parameter vector for a continuous multivariate distribution is considered. The general form of test statistic is computed by likelihood ratio method. Also, the asymptotic null distribution of test statistic is derived by limit theorems according to the chi-square distribution and critical values of test statistic for different significance levels computed and power of test estimated by using Monte Carlo simulation. The numerical examples associated with the problem of testing are presented. All the results of this paper are for independently and identically distributed random vectors. Also, the results are established for a continuous univariate distribution.