pcoscaled function

Simplified Analysis in Principal Coordinates

performs a simplified analysis in principal coordinates, using an object of class dist.

pcoscaled(distmat, tol = 1e-07)

Arguments

• distmat: an object of class dist
• tol: a tolerance threshold, an eigenvalue is considered as positive if it is larger than -tol*lambda1 where lambda1 is the largest eigenvalue

Returns

returns a data frame containing the Euclidean representation of the distance matrix with a total inertia equal to 1

References

Gower, J. C. (1966) Some distance properties of latent root and vector methods used in multivariate analysis. Biometrika, 53 , 325--338.

Author(s)

Daniel Chessel

Examples

a <- 1 / sqrt(3) - 0.2
    w <- matrix(c(0,0.8,0.8,a,0.8,0,0.8,a,
        0.8,0.8,0,a,a,a,a,0),4,4)
    w <- as.dist(w)
    w <- cailliez(w)
    w
    pcoscaled(w)
    dist(pcoscaled(w)) # w
    dist(pcoscaled(2 * w)) # the same
    sum(pcoscaled(w)^2) # unity