model_fit_glmnet: Fit glmnet models for the cbenefit package

In tbalan/cbenefit: Cross Validated C-Index and C-Index for Benefit

Description

Fit glmnet models for the cbenefit package

Usage

model_fit_glmnet(formula, data, model, pen = NULL, treat_name)

Arguments

formula	A formula to be evaluated for the model
data	A data frame or the name of the data frame where the formula may be evaluated
model	One of "coxph" for a Cox model or "glm" for logistic regression
pen	One of "ridge" or "lasso"
treat_name	Name of the treatment column as string

Details

A model of class c_glmnet is a named list where: x is the model matrix, x_other is the model matrix with the treatment allocation reversed, y is the outcome, alpha is the glmnet penalization parameter (0 for ridge, 1 for lasso) and glmnetfit is the result of glmnet::cv.glmnet. This augmented object is needed for further calculations in the cbenefit package.

The function does not spend time checking its input. In particular, the function has only been tested with complete case data frames, and where the treatment is coded as 0 (control) and 1 (treated).

Value

An object of type c_glmnet

See Also

model_fit

Examples

model_fit_glmnet(Surv(time, status) ~ treat + x1 + x2 + x3,
dat_toy, model = "coxph", pen = "ridge", treat_name = "treat")

tbalan/cbenefit documentation built on May 8, 2019, 12:58 p.m.