Fit a linear model via penalized nonconvex loss function.
ncl_fit(x,y, weights, offset=NULL,
rfamily=c("clossR", "closs", "gloss", "qloss"),
s=NULL, fk=NULL, iter=10, reltol=1e-5, trace=FALSE)An object with S3 class "ncl" for the various types of models.
the call that produced the model fit
predicted values
pseudo response values in the MM algorithm
input matrix, of dimension nobs x nvars; each row is an observation vector.
response variable. Quantitative for rfamily="clossR" and -1/1 for classifications.
observation weights. Can be total counts if responses are proportion matrices. Default is 1 for each observation
this can be used to specify an a priori known component to be included in the linear predictor during fitting. This should be NULL or a numeric vector of length equal to the number of cases. Currently only one offset term can be included in the formula.
Response type and relevant loss functions (see above)
nonconvex loss tuning parameter for robust regression and classification.
predicted values at an iteration in the MM algorithm
number of iteration in the MM algorithm
convergency criteria
If TRUE, fitting progress is reported
Zhu Wang <zwang145@uthsc.edu>
The robust linear model is fit by majorization-minimization along with least squares. Note that the objective function is $$weights*loss$$.
Zhu Wang (2021), MM for Penalized Estimation, TEST, tools:::Rd_expr_doi("10.1007/s11749-021-00770-2")
ncl