"coef"(object)
"coef"(object)
"loadings"(x, matrix = "PP", standardized = TRUE)
"loadings"(x, matrix = "PP", standardized = TRUE, level = 1)
"loadings"(x, standardized = TRUE, level = 1)
"cormat"(object, matrix = "PF")
"cormat"(object, matrix = "PF", level = 1)
"cormat"(object)
"cormat"(object, level = 1)
"uniquenesses"(object, standardized = TRUE)
"uniquenesses"(object, standardized = TRUE, level = 1)
"uniquenesses"(object, standardized = TRUE)FA-class or restrictions-classFA-class or restrictions-classloadings outputs a matrix of S3 class "loadings", which has a special
print method (see print.loadings). coef returns the primary pattern
matrix at level one and is not of class "loadings". The cormat methods
output a (symmetric) matrix, and the uniquenesses methods output a non-negative
numeric vector.
FA-class
and virtually all flavors of restrictions-class. Also,
in the code of cormat, there is a method for objects that inherit
from manifest-class.loadings methods extract the estimate of $beta$,
the cormat methods extract the estimate of $Phi$, and the
uniquenesses methods extract the diagonal of $Theta$. In addition,
the coef methods and the loadings methods that are defined for objects
restrictions-class extract the primary pattern matrix (at level 1). At the moment there is no special function to get the diagonal of $Omega$, which
is a diagonal matrix of estimated standard deviations of the manifest variables. However,
they can be extracted from the appropriate slot using the @ operator. Also, if
standardized = FALSE in the call to loadings or uniquenesses, then the
loadings or uniquenesses are scaled by these estimated standard deviations to produce
estimates on the covariance metric.
Additionally, for the loadings and cormat methods that are defined on objects of
FA-class, the matrix argument can be specified to extract a different
set of estimated coefficients or correlations. By default, matrix = "PP" for these
loadings methods, indicating that the primary pattern matrix should be extracted.
Other possible choices are "PS" to extract the primary structure matrix (defined as
$beta Phi$), "RS" to extract the reference structure matrix (which is
column-wise proportional to $beta$), "RP" to extract the reference pattern
matrix (which is column-wise proportional to $beta Phi$), and "FC"
to extract the factor contribution matrix (which is defined as
$beta * (beta Phi)$, where the * indicates element-by-element
multiplication of two matrices with the same dimensions).
By default, matrix = "PF" for these cormat methods, indicating that the
correlation matrix among primary factors should be extracted. Other possible choices
are "RF" to extract the correlation matrix among reference factors and
"PR" to extract the (diagonal) correlation matrix between primary and reference
factors.
In the case of a two-level model, the level argument can be specified to
extract such matrices from the second level of the model (including the methods
for the uniquenesses generic).
## See the example for Factanal()
Run the code above in your browser using DataLab