CRCSplineFitter

class lifelines.fitters.crc_spline_fitter.CRCSplineFitter(n_baseline_knots: int | None = None, knots: list | None = None, *args, **kwargs)

Below is an implementation of Crowther, Royston, Clements AFT cubic spline models. Internally, lifelines uses this for survival model probability calibration, but it can also be useful for a highly flexible AFT model.

Parameters:

n_baseline_knots (int) – the number of knots in the cubic spline.

References

Crowther MJ, Royston P, Clements M. A flexible parametric accelerated failure time model.

Examples

from lifelines import datasets, CRCSplineFitter
rossi = datasets.load_rossi()

regressors = {"beta_": "age + C(fin)", "gamma0_": "1", "gamma1_": "1", "gamma2_": "1"}
crc = CRCSplineFitter(n_baseline_knots=3).fit(rossi, "week", "arrest", regressors=regressors)
crc.print_summary()
fit_intercept = True
set_knots(T, E)