Why factor analysis are only applied to numeric data

r
factor_analysis

#1

I am trying to do factor analysis on one data problem but when I am applying the factor analysis model I am getting the error .

library(lattice)
my.wines <- read.csv(“http://steviep42.bitbucket.org/YOUTUBE.DIR/wines.csv”, header=TRUE)
my.prc <- prcomp(my.wines[,-1], center=TRUE, scale=TRUE)
m1_f1<-factanal(my.wines,factor=2,rotation=“varimax”)
Error in factanal(my.wines, factor = 2, rotation = “varimax”) :
factor analysis applies only to numerical variables


#2

Hi @hinduja1234,

My understanding is that mostly all dimensionality reduction techniques like PCA,SVD & factor analysis calculate eigen vectors and eigen values based on distance measures between variables.
Thus the more number of data points or values a variable has the more easy it is to calculate a distance measure among them.The concept of distance between good-bad,default-no-default etc. does not sound very scientific.
This is my general understanding,I am not sure of the exact algebra behind it.