normalize_fa {lazy.fa} | R Documentation |
Normalization of F and A in PCA of FA model
normalize_fa(F, A)
F |
n x ndim matrix of factor scores |
A |
nvar x ndim matrix of factor loadings |
The resulting matrices satisfy:
t(F)%*%F = n I, and t(A)%*%A = diagonal
A list of normalizd F
and A
seed <- 1701 set.seed(seed) n <- 200; nvar <- 9 ndim <- 3 F <- matrix(rnorm(n*ndim),n,ndim) F <- scale( F, center=TRUE ) A <- matrix(rnorm(nvar*ndim), nvar,ndim) Y <- F%*%t(A) rss <- ssq(Y-F%*%t(A)) stat=mandd( F ) Print(rss, stat$cov) Print(t(F)%*%F, t(A)%*%A, fmt="8.4") FA <- normalize_fa( F, A ) F1 <- FA$F; A1 <- FA$A rss1 <- ssq(Y-F1%*%t(A1)) stat1=mandd( F1 ) Print(rss1, stat1$cov) Print(t(F1)%*%F1, t(A1)%*%A1, fmt="8.4")