Principal Component Analysis
1. linear decomposion: SVD
1.standardized data=correlation matrix
2.principal=factor
3.not rotation
4.factor loading=correlation(z,f)
I think there are two points about factor analysis and one point about principal component analysis.
Factor analysis:
1.replicated correlations from factor loading cross products sums between variables.
2.measurement error variance/item reliabilities/commonalities from variable’s factor loading squared sum
Principal component analysis:
1.Principal componetns are the linear combinarions of original variables based eigen vectors and explain the original variables’ variances.