bda (version 18.3.2)

bootkde: Density estimation for data with rounding errors

Description

To estimate density function based on data with rounding errors.

Usage

bootkde(x,freq,a,b,from, to, gridsize=512L,method='boot')

Value

y

Estimated values of the smooth function over a fine grid.

x

grid points where the smoothed function are evaluated.

pars

return 'bw' for Berkson method, or return the mean and SD for MLE method

Arguments

x,freq

raw data if 'freq' missing, otherwise as distinct values with frequencies.

a,b

the lower and upper bounds of the rounding error.

from,to,gridsize

start point, end point and size of a fine grid where the distribution will be evaluated.

method

type estimate: "mle" or "Berkson", or "boot" (default).

References

Wang, B and Wertelecki, W, (2013) Computational Statistics and Data Analysis, 65: 4-12.

Examples

 #data(ofc)
 #x0 = round(ofc$Head)
 x0 = round(rnorm(100,34.5,1.6))
 fx1 = bootkde(x0,a=-0.5,b=0.5,method="Berkson")