icenReg (version 2.0.16)

diag_baseline: Compare parametric baseline distributions with semi-parametric baseline

Description

Creates plots to diagnosis fit of different choices of parametric baseline model. Plots the semi paramtric model against different choices of parametric models.

Usage

diag_baseline(
  object,
  data,
  model = "ph",
  weights = NULL,
  dists = c("exponential", "weibull", "gamma", "lnorm", "loglogistic", "generalgamma"),
  cols = NULL,
  lgdLocation = "bottomleft",
  useMidCovars = T
)

Arguments

object

Either a formula or a model fit with ic_sp or ic_par

data

Data. Unnecessary if object is a fit

model

Type of model. Choices are 'ph' or 'po'

weights

Case weights

dists

Parametric baseline fits

cols

Colors of baseline distributions

lgdLocation

Where legend will be placed. See ?legend for more details

useMidCovars

Should the distribution plotted be for covariates = mean values instead of 0

Author

Clifford Anderson-Bergman

Details

If useMidCovars = T, then the survival curves plotted are for fits with the mean covariate value, rather than 0. This is because often the baseline distribution (i.e. with all covariates = 0) will be far away from the majority of the data.

Examples

data(IR_diabetes)
fit <- ic_par(cbind(left, right) ~ gender, 
             data = IR_diabetes)

diag_baseline(fit, lgdLocation = "topright", 
             dist = c("exponential", "weibull", "loglogistic"))