Question

I am doing SVD in R on a data frame called data

svd1 <- svd(scale(data))

I plot the result using

plot(svd1$d^2/sum(svd1$d^2),xlab="Column",ylab="Percent of variance explained",pch=19)

I found out that in the plot, column number do not corespond to the column numbers in the data frame (no matter what subset of columns I use with SVD, the first column always shows the highes variance).

My question is, how do I get the column names (or “real” indices) in the plot?

Answer 1

According to the R documentation (…/library/base/html/svd.html):

d is a vector containing the singular values of x, of length min(n, p).

These singular values are computed by the SVD algorithm from the whole input matrix, so there is no way to label the singular values based on the column names.