Kernel Estimation of Hazard Function in Survival Analysis

Producing kernel estimates of the unconditional and conditional hazard function for right-censored data including methods of bandwidth selection.

kernhaz

The goal of kernhaz is to estimate the unconditional and the conditional hazard function for right-censored data. The package includes two methods of the bandwidth selection - crossvalidation and maximum likelihood method.

Installation

You can install the released version of kernhaz from CRAN with:

install.packages("kernhaz")

Example

The unconditional hazard function, the bandwidth is selected using the crossvalidation method:

library(survival)
fit1<-khazard(times = lung$time,delta = lung$status-1)
plot(fit1)

The unconditional hazard function, four bandwidths are considered:

library(survival)
fit2<-khazard(times = lung$time,delta = lung$status-1,h=c(100,150,200,250), value="both")
plot(fit2,h=200)

The conditional hazard function, bandwidths are selected using the crossvalidation method:

library(survival)
fit3<-khazardcond(times = lung$time,delta = lung$status-1,covariate = lung$age)
plot(fit3)

The conditional hazard function, bandwidths are user's choice:

library(survival)
fit4<-khazardcond(times = lung$time,delta = lung$status-1,covariate = lung$age,h=c(200,20))
plot(fit4)