Minimises factanal's objective over uniquenesses between 0.005 and 1 by projected BFGS
steps; returns unrotated loadings, uniquenesses, Bartlett's chi-square, AIC and BIC.
Usage
MlFac(X, n_factors = NULL, max_iter = 500, tol = 1e-12, scale = TRUE)
Arguments
- X
Data matrix.
- n_factors
Number of factors (default: Kaiser count).
- max_iter
BFGS iterations.
- tol
Gradient tolerance.
- scale
Analyse correlations (TRUE) or covariances.
Value
list(loadings, communalities, uniqueness, variance_explained, objective,
log_likelihood, statistic, dof, p_value, aic, bic, n_factors).
References
Joreskog, K. G. (1967). Psychometrika 32, 443-482.
Examples
X <- cbind(1:8, c(2, 3, 3, 5, 5, 7, 8, 8), c(1, 2, 4, 4, 6, 5, 7, 9), c(3, 4, 4, 6, 5, 8, 8, 9))
MlFac(X, 1)$uniqueness
#> [1] 0.00500000 0.04013841 0.08665601 0.08274747
