Survival analysis deals with time to an event in systems. Events can be death in biological systems and failure in technical systems, but the event may be something entirely different, such as divorce, relapse of a disease, or an insurance claim. Often the time to an event is not known exactly, but is known to fall in some interval; this is called censoring. The Wolfram Language allows you to specify time-to-event data in a flexible (censor intervals, indicators, or counts) and powerful (right, left, interval censoring, and truncation) way. Time-to-event data is broadly supported throughout the system. Time-to-event data can be used to compute descriptive statistics, estimate parametric and nonparametric distributions, fit a variety of survival models, and perform hypothesis tests.

Survival Data

EventData — right, left, interval-censored, and truncated data

Descriptive Survival Statistics »

Median — median life of survival data and distributions

Quantile   ▪ Quartiles   ▪ InterquartileRange   ▪ Mean   ▪ ...

Nonparametric Survival Estimators »

SurvivalModelFit — distribution with confidence intervals (Kaplan–Meier, …)

EmpiricalDistribution   ▪ SmoothKernelDistribution   ▪ ...

Parametric Survival Estimators »

EstimatedDistribution — estimate parametric distribution from survival data

WeibullDistribution   ▪ ExponentialDistribution   ▪ GompertzMakehamDistribution   ▪ LogNormalDistribution   ▪ ...

Proportional Hazards Modeling

CoxModelFit — estimate a Cox proportional hazards model

StrataVariables   ▪ NominalVariables   ▪ ConfidenceTransform

Hypothesis Tests for Survival Data

LogRankTest — test whether hazard rates are equivalent

Related Guides

▪
• Reliability Analysis
▪
• Nonparametric Distributions
▪
• Probability & Statistics
▪
• Regression Modeling