Development build. This page describes main, not a released package. The latest published Lodestar.Survival is 0.2.0 — read its documentation.

Home › Survival › Survival estimators

Each subject's survival at every event time: lifelines' predict_survival_function.

public double[] PredictSurvivalFunction(ReadOnlySpan<double> design)

Parameters — design holds the subjects' covariates row-major, FeatureCount per row; empty for the one subject of a fit with no covariate.

Returns — row-major, one row per subject and one column per EventTimes entry: exp(−H) of the cumulative hazard.

Exceptions — ArgumentException when design is not whole rows of finite covariates. ArgumentException naming design also when it, with any intercept column, or its rows times the event times, is more cells than one array holds (#1614, #1616).

Example — an untreated and a treated patient, at the last of the ten event times.

using Lodestar.Survival;

double[] treated = [0, 1, 0, 1, 0, 1, 0, 1, 0, 1, 0, 1];
double[] months = [3, 10, 5, 4, 8, 18, 12, 6, 9, 20, 7, 15];
bool[] died = [true, true, true, false, true, true, true, true, true, false, true, true];

AalenSummary fit = AalenAdditive.Fit(treated, months, died, featureCount: 1);

double[] survival = fit.PredictSurvivalFunction([0.0, 1.0]);
double untreated = Math.Round(survival[9], 6);   // => 0.086294
double onTreatment = Math.Round(survival[19], 6);   // => 0.693041

Remarks — the exponential of PredictCumulativeHazard, read only at the event times. The curve need not fall monotonically: an increment may be negative, and survival then rises, past one where the cumulative hazard goes below zero.

Applies to — net10.0, netstandard2.0.

See also — AalenSummary, AalenSummary.PredictPercentile.