Skip to content
Prev 177874 / 398502 Next

search through a matrix

I changed the condition to:
 highlyCor<-cbind(rownames(x)[row(x)[((x > 0.9)|(x<(-.9)))]],
colnames(x)[col(x)[((x > 0.9)|(x<(-.9)))]]) 

Actually what I am trying to do is run factor analysis on this 923 by 925 x
matrix but many of the variables where too highly correlated and so I got
some message about singular and that the factanal could not run. 
Is there a way to choose to remove one variable from each pair of highly
correlated variables (say variables with greater than .9 or less than -.9
correlation) and receive a new smaller x matrix which I can then do factor
analysis on?

Thanks.
Uwe Ligges-3 wrote: