R/cor2m.R

In phiala/ecodist: Dissimilarity-Based Functions for Ecological Analysis

# Correlate variables from 1 matrix with variables from another
# matrix. If TRIM, set rho<critical value(alpha) to NA. Computes
# this critical value as a t-test with n-2 df.

cor2m <- function(x, y, trim=TRUE, alpha=0.05) {
    xz <- scale(as.matrix(x))
    yz <- scale(as.matrix(y))
    n <- dim(x)[[1]]
    cc <- t(yz) %*% xz / (n -1.0)
    rownames(cc) <- colnames(y)
    colnames(cc) <- colnames(x)
    if (trim) {
        rt <- cc * sqrt((n-2)/(1-cc^2))
        cv <- qt(1-alpha, n-2)
        cc[abs(rt) < cv] <- NA
    }
    data.frame(t(cc))
}