R 函数获取带有协变量的 Kendall tau p 值？

Question

我有两个变量和第三个协变量调整器。如何获得调整后的 Kendall tau 估计值加上一个 p 值？

a <- c(1.07, 1.9, -0.603, -0.391, -0.416, -0.376, -0.367, -0.296, 1.44, -0.698) 
b <- c(1.13, 1.95, 0.37, 0.404, -0.385, 0.168, -0.349, 0.481, 2.2, -0.687)
c <- c(3.75, 3.75, 3.74, 3.75, 3.75, 3.74, 3.74, 5.37, 8.18, 8.18)

我希望获得与

cor.test(a, b, method = "kendall")

类似的输出，该输出已针对

c

进行了调整。

Answer 1

这是一个运行偏相关测试的函数

pcor.test

。该函数基于函数

pcor

，包

ppcor

。

pcor.test <- function (x, y, z, method = c("pearson", "kendall", "spearman"), conf.level = 0.95, ...) {
  d1 <- deparse(substitute(x))
  d2 <- deparse(substitute(y))
  d3 <- deparse(substitute(z))
  data.name <- paste0(d1, " and ", d2, "; controlling: ", d3)
  method <- match.arg(method)
  Method <- paste0("Partial correlation (", method, ")")
  alternative <- "true partial correlation is not equal to 0"
  
  x <- as.vector(x)
  y <- as.vector(y)
  z <- as.data.frame(z)
  xyz <- data.frame(x, y, z)
  pc <- ppcor::pcor(xyz, method = method)
  estimate <- pc$est[1, 2]
  p.value <- pc$p.value[1, 2]
  parameter <- c(n = pc$n, gp = pc$gp)
  statistic <- c(Stat = pc$statistic[1, 2])
  
  ht <- list(
    statistic = statistic,
    parameter = parameter,
    p.value = p.value,
    estimate = c(partial.cor = estimate),
    alternative = alternative,
    method = Method,
    data.name = data.name
  )
  class(ht) <- "htest"
  ht
}

a <- c(1.07, 1.9, -0.603, -0.391, -0.416, -0.376, -0.367, -0.296, 1.44, -0.698) 
b <- c(1.13, 1.95, 0.37, 0.404, -0.385, 0.168, -0.349, 0.481, 2.2, -0.687)
c <- c(3.75, 3.75, 3.74, 3.75, 3.75, 3.74, 3.74, 5.37, 8.18, 8.18)
pcor.test(a, b, c, method = "kendall")
#> 
#>  Partial correlation (kendall)
#> 
#> data:  a and b; controlling: c
#> Stat = 2.5651, n = 10, gp = 1, p-value = 0.01032
#> alternative hypothesis: true partial correlation is not equal to 0
#> sample estimates:
#> partial.cor 
#>   0.6834271

^{创建于 2024-04-19，使用 reprex v2.1.0}