Description Usage Arguments Details Value Methods Author(s) Examples
These functions extract various elements of formal S4 objects that are important in factor analysis models, namely the loadings, the correlations among factors, and the unique variances. Occasionally, it may be useful to call these generic functions directly.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 | ## S4 method for signature 'FA'
coef(object)
## S4 method for signature 'restrictions'
coef(object)
## S4 method for signature 'FA'
loadings(x, matrix = "PP", standardized = TRUE)
## S4 method for signature 'FA.general'
loadings(x, matrix = "PP", standardized = TRUE, level = 1)
## S4 method for signature 'restrictions.general'
loadings(x, standardized = TRUE, level = 1)
## S4 method for signature 'FA'
cormat(object, matrix = "PF")
## S4 method for signature 'FA.2ndorder'
cormat(object, matrix = "PF", level = 1)
## S4 method for signature 'restrictions'
cormat(object)
## S4 method for signature 'restrictions.2ndorder'
cormat(object, level = 1)
## S4 method for signature 'FA'
uniquenesses(object, standardized = TRUE)
## S4 method for signature 'FA.general'
uniquenesses(object, standardized = TRUE, level = 1)
## S4 method for signature 'restrictions'
uniquenesses(object, standardized = TRUE)
|
object |
an object that inherits from |
x |
an object that inherits from |
matrix |
a character string with exactly two letters indicating which matrix to extract; see the Details section |
standardized |
a logical indicating whether to standardize the result so that it is calibrated for a correlation matrix among manifest variables, rather than their covariance matrix |
level |
either 1 or 2 to indicate from which level of the factor analysis model is pertinent when the model has two levels |
Let the factor analysis model be
Sigma = Omega(beta Phi beta' + Theta)Omega
By default, the 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).
loadings
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.
There are methods for every flavor of 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
.
Ben Goodrich
1 | ## See the example for Factanal()
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.