Bootstrap Tests For The Equality Of Location Parameter In Two And More Multi-dimension Samples

In 1998, Shiraishi T. used M-statistics to test the equality of location parameter in the model of two one-dimensional samples and contrasted the results with t-test. This dissertation gives two P.P. type M-statistics and one P.P. type t-statistic to test the equality of location parameter in two and more multi-dimension samples, respectively. We cannot get the accurate distribution of the statistics above for the population distribution is not known. It will be used bootstrap method in this dissertation, and two P.P. type bootstrap M-statistics and one P.P. type bootstrap t-statistic to approximate the distributions of the statistics above, respectively. We will do a lot of simulations to test and simulate the statistics above and to test which one is better; moreover we will give and analyses the results of the tests and compare the two P.P. type M-statistics, the P.P. type t-statistic with each other.This dissertation will be made up of four parts. Chapter 1 is preparatory knowledge; and we will introduce bootstrap tests for the equality of location parameter in two, more multi-dimension samples in Chapter 2, Chapter 3, respectively. The conclusion and the creative aspect will be elaborated in Chapter 4.