bwe.cox: Backward feature elimination based on Wald type test for Cox...

In linxihui/sml: Machine Learning for Survival Data

Description

Backward feature elimination based on Wald type test for Cox model

Usage

bwe.cox(formula, data, alpha = 0.05, trace = 1, method = "breslow", ...)

Arguments

formula	A model formula to start backward elimination
data	A data frame
alpha	Significant level. Default to 0.05
trace	A integer for different levels of tracing. 0 = no trace, 1 = print dropped variable, 2 = print all considered models during the procedure
method	Method to resolve tied event times pass to ‘coxph'. Default to ’breslow' instead of 'efron'
...	Other parameters passed 'coxph'

Details

Notice that model hierachical structure is maintained, i.e., iff interactions involved in the model, its main effects will also be kept in the model.

Value

A list with two entries:
$ fit: a 'coxph' object of selected model $ wald.test: wald test p values for each feature and possible interaction.

linxihui/sml documentation built on May 21, 2019, 6:39 a.m.