Hazard Function Estimation in Survival Analysis

Produces a smooth estimate of the hazard function for censored data.

muhaz

Smooth hazard function estimation from right-censored survival data.

Overview

muhaz estimates the hazard function from right-censored data using kernel-based methods, implementing the bandwidth selection algorithms and boundary kernel formulations described in Mueller and Wang (1994). Options include:

• Three bandwidth methods: local MSE-optimal, global IMSE-optimal, and k-nearest-neighbor
• Three boundary correction types: none, left only, or both ends
• Four kernel shapes: rectangle, Epanechnikov, biquadratic, triquadratic

A complementary set of piecewise-exponential estimators (pehaz, plot.pehaz) is also provided for quick exploratory comparison.

Original S code by Kenneth Hess (M.D. Anderson Cancer Center); R port by R. Gentleman. Currently maintained by David Winsemius.

Installation

Install the released version from CRAN:

install.packages("muhaz")

Install the development version from GitHub:

# install.packages("remotes")
remotes::install_github("dwinsemius/muhaz")

Quick start

library(muhaz)
data(cancer, package = "survival")

# Locally optimal bandwidth (default)
fit <- muhaz(ovarian$futime, ovarian$fustat)
plot(fit)
summary(fit)

# Globally optimal bandwidth
fit_global <- muhaz(ovarian$futime, ovarian$fustat, bw.method = "g")

# Fixed bandwidth
fit_fixed <- muhaz(ovarian$futime, ovarian$fustat, bw.method = "g", bw.grid = 5)

References

1. Mueller HG, Wang JL. Hazard rates estimation under random censoring with varying kernels and bandwidths. Biometrics 1994; 50: 61--76.

2. Gefeller O, Dette H. Nearest neighbour kernel estimation of the hazard function from censored data. J Statist Comput Simul 1992; 43: 93--101.

3. Hess KR, Serachitopol DM, Brown BW. Hazard function estimators: a simulation study. Statistics in Medicine 1999.

License

GPL. See COPYING for details.