corr: 2nd momments of sparse matrices
In textir: Inverse Regression for Text Analysis

Description Usage Arguments Value Author(s) See Also Examples

View source: R/dtm.R

Correlation and deviation in sparse matrices.

1 2	corr(x, y) sdev(x)

`x`	A `dgCMatrix` or `matrix` of counts.
`y`	A `matrix` with `nrow(y)=nrow(x)`.

corr returns the ncol(x) by ncol(y) matrix of correlation between x and y, and sdev returns the column standard deviations.

Matt Taddy taddy@chicagobooth.edu

pls, congress109

# some congress examples
data(congress109)
r <- corr(congress109Counts, congress109Ideology$repshare)
## 20 terms for Democrats
sort(r[,1])[1:20]
## 20 terms for Republicans
sort(r[,1], decreasing=TRUE)[1:20]
## 20 high variance terms
colnames(congress109Counts)[
	order(-sdev(congress109Counts))[1:20]]