widenet: Penalized Regression with Polynomial Basis Expansions

Extends the glmnet and relaxnet packages with polynomial basis expansions. Basis expansion is applied to the predictors and a subset of the basis functions is chosen using relaxnet. Predictors may be screened using correlation or t-tests. Screening is done separately within cross-validation folds. Cross-validation may be used to select the order of basis expansion and alpha, the elastic net tuning parameter.

Author
Stephan Ritter, Alan Hubbard
Date of publication
2013-07-29 21:30:08
Maintainer
Stephan Ritter <sritter@berkeley.edu>
License
GPL (>= 2)
Version
0.1-2
URLs

View on CRAN

Man pages

predict.widenet
Predict Method for '"widenet"' Objects
print.widenet
Print Method for widenet Objects
summary.widenet
Generate and print summaries of class '"widenet"' objects.
widenet
Extends the relaxnet Package with Polynomial Basis Expansions

Files in this package

widenet
widenet/src
widenet/src/expandBasis.c
widenet/NAMESPACE
widenet/R
widenet/R/screen.R
widenet/R/checkBinary.R
widenet/R/summary.widenet.R
widenet/R/expandBasis.R
widenet/R/widenet.cv.funs.R
widenet/R/predict.widenet.R
widenet/R/zzz.R
widenet/R/widenet.R
widenet/MD5
widenet/DESCRIPTION
widenet/man
widenet/man/print.widenet.Rd
widenet/man/summary.widenet.Rd
widenet/man/predict.widenet.Rd
widenet/man/widenet.Rd