不好意思,是我的理解错误,将你的偏相关系数理解成了偏自相关系数.

这是两个不同的概念.

发现这个问题之后, 我也在想如何计算偏相关系数, 找了好的资料,呵呵都没有讲这个问题.

发现SPLUS 或R 中没有现成的函数可以计算它们.

但是还是有一些人自己写code来计算. 在R中

John Fox 的代码:

Suppose that R is the correlation matrix among the variables. Then

Rinv <- solve(R)
D <- diag(1 / sqrt(diag(Rinv)))
P <- -D %*% Rinv %*% D

Martyn Plummer:

In general you invert the variance-covariance matrix and then rescale it so the diagonal is one. The off-diagonal elements are the negative partial correlation coefficients given all other variables.

pcor2 <- function(x){
conc <- solve(var(x))
resid.sd <- 1 / sqrt(diag(conc))
pcc <- -sweep(sweep(conc, 1, resid.sd, "*"), 2, resid.sd, "*")
return(pcc)
}
pcor2(cbind(x1, x2, x3))

KP responded: This is the version I’m using now, together with a test for significance of each coefficient (H0: coeff = 0):

f.parcor <- function (x, test = FALSE, p = 0.05) {
nvar <- ncol(x)
ndata <- nrow(x)
conc <- solve(cor(x))
resid.sd <- 1 / sqrt(diag(conc))
pcc <- -sweep(sweep(conc, 1, resid.sd, "*"), 2, resid.sd, "*")
colnames(pcc) <- rownames(pcc) <- colnames(x)
if (test) {

t.df <- ndata - nvar
t <- pcc / sqrt((1 - pcc^2) / t.df)
pcc <- list(coefs = pcc, significant = t > qt(1 - (p/2),
df = t.df))
}
return(pcc)
}