A New High-dimensional Two-sample Mean Test

In today's era,with the rise of genetic problems,a new kind of data,high-dimensional data,appears in various fields,which brings great challenges to the traditional statistical multivariate statistical theory and related methods.In the hypothesis testing brought by high-dimensional data,on the one hand,the classical test statistics are meaningless in the case of high-dimensional data;On the other hand,when the sample size n is far larger than the dimension p,the limit theory is no longer applicable.Creating limit theory suitable for high-dimensional data and improving the original statistics are popular areas of research nowadays.Based on the hypothesis test of the population mean,this article puts forward new ideas and new methods.Based on the statistical bootstrap method,the optimal statistics under certain conditions are proposed by improving the original statistics.At the same time,through the data simulation in R language,it is found that compared with previous methods,the statistics we put forward are superior in power and p-value under certain conditions,and the statistical bootstrap method we use also avoids the difficulty of theoretical calculation.