factor retention methods

[mdsteiner]

To implement:

• use EFA (also ML and ULS possibility) instead of only PAF in HULL and PARALLEL
• use ... for additional arguments to EFA in HULL, PARALLEL; remove arguments "criterion", "criterion_type", and "max_iter"

--> done

[sgrieder]

N_FACTORS function:

• reference to function in all the separate fac retention functions --> done

• Check for singular matrix only under certain conditions (e.g., only if certain tests are performed)? --> no, check always

[sgrieder]

Update PARALLEL:

• Check literature: Revised PA by Green et al 2012 for eigenvalues of EFA and 95th percentile as reference eigenvalue --> in Auerswald not superior to other methods, not implemented

[sgrieder]

On deviance vs. chi square for AIC:

• https://online.stat.psu.edu/stat504/node/60/