fit.lognet: Tailormade weighted fitting of logistic glmnet

In EMLasso: Fit (logistic) LASSO when there is missing data in the predictors

Description

Do dummy coding on dataset and fit logistic (crossvalidated) glmnet

Usage

  fit.lognet(ds, out, lambda,
    weights = (rep(1, dim(ds)[1])), verbosity = 0,
    standardize = FALSE, type.measure = NULL,
    imputeDs2FitDsProperties = normalImputationConversion(),
    ...)

Arguments

ds	dataset (numdfr or data.frame)
out	outcome vector
lambda	(single) lambda to use
weights	weight vector per observation (does not have to sum to 1, and defaults to equal weights)
verbosity	The higher this value, the more levels of progress and debug information is displayed (note: in R for Windows, turn off buffered output)
standardize	if TRUE (NOT the default), all variables are standardized before fitting.
type.measure	one of the crossvalidating measures provided by glmnet or NULL (default) to not do crossvalidation
imputeDs2FitDsProperties	see imputeDs2FitDs and EMLasso
...	passed on to glmnet or cv.glmnet.

Value

depending on type.measure being NULL, a glmnet or cv.glmnet object.

Author(s)

Nick Sabbe [e-mail]

See Also

glmnet

Examples

y<-rbinom(nrow(iris), 1, 0.5)
lnet<-fit.lognet(iris, y, lambda=0.05, weights=runif(nrow(iris)), verbosity=1)
cv.lnet<-fit.lognet(iris, y, lambda=0.05, weights=runif(nrow(iris)), verbosity=1, type.measure="auc")

EMLasso documentation built on May 2, 2019, 5:49 p.m.