BackSelection: Build the model through backward selection

In ywang297/CoxFlex: Flexible Modeling of the Time-to-Event Data

Build the model through backward selection

Description

Build the model through backward selection

Usage

BackSelection(data, Type, variables, continuous, TD, NL, m = 1, p = 2,
  alpha_back = 0.05, knots = -999)

Arguments

data	A data frame in the long (interval) format with one line per unit of time.
Type	A vector consisting the name of variables representing the start and stop of each time interval and event indicator. e.g. c("start","stop","event").
variables	A vector consisting the name of varialbes that will be adjusted in the model.
continuous	A vector of binary indicators representing whether the corresponding variables are continuous variables (1=Yes, 0=NO)
TD	A vector representing whether to force the corresponding variable having a time-dependent effect. Only takes three values: 1 = force TD effect of the corresponding variable 0 = do not force any effect to the corresponding variable -1 = force the PH effect of the corresponding variable
NL	A vector representing whether to force the corresponding variable having a non-linear effect. Only takes three values: 1 = force NL effect of the corresponding variable 0 = do not force any effect to the corresponding variable -1 = force the linear effect of the corresponding variable
m	The number of interior knots used in the regression B-spline, default value is 1
p	The degree of the regression B-spline, default value is 2
alpha_back	significane level used in the backward elemination process
knots	Default value is -999

Note

Note that the TD and NL arguments have different meanings than that in the CoxFlex and tvcFlex

ywang297/CoxFlex documentation built on March 18, 2024, 3:26 a.m.