muhaz (version 1.2.6.1)

muhaz: Estimate hazard function from right-censored data.

Description

Estimates the hazard function from right-censored data using kernel-based methods. Options include three types of bandwidth functions, three types of boundary correction, and four shapes for the kernel function. Uses the global and local bandwidth selection algorithms and the boundary kernel formulations described in Mueller and Wang (1994). The nearest neighbor bandwidth formulation is based on that described in Gefeller and Dette (1992). The statistical properties of many of these estimators are reported and compared in Hess et al (1999). Based on the HADES program developed by H.G. Mueller. Returns an object of class 'muhaz.' NOTE: For comparison to the smoothed hazard function estimates, we have also made available a set of functions based on piecewise exponential estimation. These estimates are similar in concept to the histogram estimator of the density function. They give a feel for the features of the data without the manipulations involved in smoothing. They also help to confirm that muhaz is generating realistic estimates of the underlying hazard function. These functions are called: pehaz, plot.pehaz, lines.pehaz, print.pehaz.

Usage

muhaz(times, delta, subset, min.time, max.time, bw.grid, bw.pilot,
      bw.smooth, bw.method="local", b.cor="both", n.min.grid=51,
      n.est.grid=101, kern="epanechnikov")

Arguments

times

Vector of survival times. Does not need to be sorted.

delta

Vector indicating censoring 0 - censored (alive) 1 - uncensored (dead) If delta is missing, all the observations are assumed uncensored.

subset

Logical vector, indicating the observations used in analysis. T - observation is used F - observation is not used If missing, all the observations will be used.

min.time

Left bound of the time domain used in analysis. If missing, min.time is considered 0.

max.time

Right bound of the time domain used in analysis. If missing, max.time is the time at which ten patients remain at risk.

bw.grid

Bandwidth grid used in the MSE minimization. If bw.method="global" and bw.grid has one component only, no MSE minimization is performed. The hazard estimates are computed for the value of bw.grid. If bw.grid is missing, then a bandwidth grid of 21 components is built, having as bounds: $$[0.2*bw.pilot, 20*bw.pilot]$$

bw.pilot

Pilot bandwidth used in the MSE minimization. If missing, the default value is the one recommended by Mueller and Wang (1994): $$bw.pilot = (max.time-min.time) / (8*nz^0.2)$$ where nz is the number of uncensored observations

bw.smooth

Bandwidth used in smoothing the local bandwidths. Not used if bw.method="global" If missing: $$bw.smooth = 5 * bw.pilot$$

bw.method

Algorithm to be used. Possible values are: "global" - same bandwidth for all grid points. The optimal bandwidth is obtained by minimizing the IMSE. "local" - different bandwidths at each grid point. The optimal bandwidth at a grid point is obtained by minimizing the local MSE. "knn" - k nearest neighbors distance bandwidth. The optimal number of neighbors is obtained by minimizing the IMSE. Default value is "local". Only the first letter needs to be given (e.g. "g", instead of "global").

b.cor

Boundary correction type. Possible values are: "none" - no boundary correction "left" - left only correction "both" - left and right corrections Default value is "both". Only the first letter needs to be given (e.g. b.cor="n").

n.min.grid

Number of points in the minimization grid. This value greatly influences the computing time. Default value is 51.

n.est.grid

Number of points in the estimation grid, where hazard estimates are computed. Default value is 101.

kern

Boundary kernel function to be used. Possible values are: "rectangle", "epanechnikov", "biquadratic", "triquadratic". Default value is "epanechnikov". Only the first letter needs to be given (e.g. kern="b").

Value

Returns an object of class 'muhaz', containing input and output values. Methods working on such an object are: plot, lines, summary. For a detailed description of its components, see object.muhaz.

Details

The muhaz object contains a list of the input data and parameter values as well as a variety of output data. The hazard function estimate is contained in the haz.est element and the corresponding time points are in est.grid. The unsmoothed local bandwidths are in bw.loc and the smoothed local bandwidths are in bw.loc.sm.

For bw.method = 'local' or 'knn', to check the shape of the bandwidth function used in the estimation, use plot(fit$pin$min.grid, fit$bw.loc) to plot the unsmoothed bandwidths and use lines(fit$est.grid, fit$bw.loc.sm) to superimpose the smoothed bandwidth function. Use bw.smooth to change the amount of smoothing used on the bandwidth function.

For bw.method='global', to check the minimization process, plot the estimated IMSE values over the bandwidth search grid. Use plot(fit$bw.grid, fit$globlmse). Use k.grid and k.imse for bw.method='k'. You may want to repeat the search using a finer grid over a shorter interval to fine-tune the optimization or if the observed minimum is at the extreme of the grid you should specify a different grid.

References

1. H.G. Mueller and J.L. Wang - Hazard Rates Estimation Under Random Censoring with Varying Kernels and Bandwidths, Biometrics 50, 61-76, March 1994

2. O. Gefeller and H. Dette - Nearest Neighbour Kernel Estimation of the Hazard Function From Censored Data, J. Statist. Comput. Simul., Vol.43, 1992, 93-101

3. K.R. Hess, D.M. Serachitopol and B.W. Brown - Hazard Function Estimators: A Simulation Study, Statistics in Medicine (in press).

See Also

summary.muhaz, plot.muhaz, lines.muhaz, muhaz.object

Examples

Run this code
# NOT RUN {
# to compute a locally optimal estimate
data(ovarian, package="survival")
attach(ovarian)
fit1 <- muhaz(futime, fustat)
plot(fit1)
summary(fit1)
# to compute a globally optimal estimate
fit2 <- muhaz(futime, fustat, bw.method="g")
# to compute an estimate with global bandwidth set to 5
fit3 <- muhaz(futime, fustat, bw.method="g", bw.grid=5)
# }

Run the code above in your browser using DataCamp Workspace