Description Objects from the Class Methods Author(s) References See Also Examples
An object containing the output of MRSP
-calls with more than one
tuning parameter value, i.e. model output related to a whole “coefficient path”.
With nrlambda
denoting the number of different lambda values for which to fit
a penalized multinomial response model, MRSP.list
-objects are lists of length
nrlambda
. Each entry of this list is an object of class MRSP
and contains the model output for one concrete lambda value. If an MRSP.list
-object
was created by function MRSP
, it additionally has
attributes “topcall”, “call” and “dat”. Those attributes store, respectively, the call to
MRSP
, the call to MRSP.fit
which was prepared by MRSP
and
the data object to be supplied to MRSP.fit
. Note that
dat
contains the standardized covariates if standardize = TRUE
.
Objects of this class are created by function MRSP.fit
, which will
most likely be called internally by the end-user function MRSP
.
signature(object = "MRSP.list")
: cv(object, k, type, parallel, cores, ...)
performs k
-fold crossvalidation of the models in object
.
The output is a list whose entry mean
gives the mean values of the criterion specified via argument type
over the crossvalidation folds; for each model found in object
.
type
can be "deviance"
(the default), "loglik"
or "brier"
, using the deviance, loglikelihood or the Brier score. Depending on the logical argument parallel
and integer argument cores
,
crossvalidation is performed either serially or with cores
parallel processes.
signature(object = "MRSP.list")
: extract(object, slotname)
returns the value of the slot whose name was given as a character string
from all elements of object
. The result is coerced to a vector or matrix if appropriate.
signature(x = "MRSP.list")
: Get the call that created the MRSP.list
-object.
signature(x = "MRSP.list")
: Plot coefficient paths.
signature(object = "MRSP.list")
: select(object, criterion=c("AIC","BIC","cv"), ...)
selects one concrete MRSP
model from object
based on criterion
.
For criterion="cv"
, an additional, integer argument k
, a logical argument parallel
and an integer argument cores
can be supplied, see cv
above.
signature(object = "MRSP.list")
: Print some basic infos about the MRSP.list
object.
Wolfgang Poessnecker
Tutz, G., Poessnecker, W., Uhlmann, L. (2015)
Variable Selection in General Multinomial Logit Models
Computational Statistics and Data Analysis, Vol. 82, 207-222.
http://www.sciencedirect.com/science/article/pii/S0167947314002709
1 | showClass("MRSP.list")
|
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.