基于T-test的permutation test

基于T-test的permutation test

1：R 中进行置换检验的方式有多种，比较典型的就是coin包中的oneway_test函数，但是一个问题是实际问题考虑的可能只是单侧的置换检验的结果，
  但是oneway_test计算的默认是双侧的P值，这时的计算最好是用Deducer包中的perm.t.test函数

   

   

2：简单的perm.t.test函数的说明

  perm.t.test(x,y,statistic=c("t","mean"),alternative=c("two.sided", "less", "greater"), midp=TRUE, B=10000)

 x    　　　　　　 a numeric vector containing the first sample
 y   　　　　　　  a numeric vector containing the second sample
 statistic    　 The statistic to be permuted. See details
 alternative     The alternative hypothesis，可以设置是单侧的还是双侧的假设检验
 midp   　　　　  should the mid p-value be used
 B    　　　　　　 The number of monte-carlo samples to be generated

 Details
 This function performs a two sample permutation test. If the mean is permuted, then the test assumes exchangability between the two samples. 
 if the t-statistic is used, the test assumes either exchangability or a sufficiently large sample size. Because there is little lost in the way of power, 
 and the assumptions are weaker, the t-statistic is used by default.

 Value
 A list with class "htest" containing the following components: 
 statistic      　　The observed value of the statistic.
 p.value        　　the p-value for the test.
 method         　　a character string indicating the type of test performed.
 data.name      　　a character string giving the name(s) of the data.
 B              　　The number of samples generated
 alternative    　　the direction of the test

 4：例子默认的是以Ttest的方式进行置换检验
 perm.t.test(rnorm(100),runif(100,-.5,.5))
 rm(list=ls())
 x<-c(1,2,3,4,5)
 y<-c(100,101,102,103,104)
 perm.t.test(y,x,alternative="greater",midp=TRUE, B=1000)

5：使用注意：
  a：不同类样本的先后顺序
  b：函数perm.t.test处于Deducer包中，在安装Deducer的过程中需要涉及java的环境配置等等，
     需要R的版本与Java的版本一致，目前找到的比较好的处理方式参照如下：http://www.cnblogs.com/ohshit/p/6159644.html
     处理之后便可以成功library rJava