Skip to contents

These functions create gamlss2 family objects for accelerated failure time models with five error distributions. Each family handles Surv() responses with right censoring, left censoring, and left truncation.

Usage

AFT_Weibull(theta = 0L)

AFT_Logistic(theta = 0L)

AFT_Normal(theta = 0L)

AFT_Cauchy(theta = 0L)

AFT_Gamma(theta = 0L, nu = 1)

Arguments

theta

Integer, order of the theta polynomial linking log(sigma) to mu. Default 0 (no coupling).

Value

A gamlss2.family object.

Details

The model is: $$\log(T) = \mu + \sigma_{eff} \cdot \varepsilon$$ where \(\log(\sigma_{eff}) = \log(\sigma) + \sum_k \theta_k (\mu - \bar\mu)^k\)

Parameters:

mu

Location parameter (identity link)

sigma

Scale parameter (log link)

theta1, theta2, ...

Coupling parameters (identity link), if theta > 0