h.crit: Bandwidth calculation for density estimation

In jenzopr/silvermantest: Investigate the Number of Modes using Kernel Density Estimates

Description

Calculates the smallest value so that the gaussian kernel density estimate of the given data x has k modes. The smaller you choose the bandwidth for a kernel density estimate, the larger the number of modes becomes. This function calculates the smallest value leading to a kernel density estimate with k number of modes.

Usage

h.crit(x, k, prec = 6, density.fun = NULL)

Arguments

x	vector of data
k	number of modes
prec	number of digits for precision of calculation
density.fun	A function that returns a vector of density estimates

Value

the smallest value so that the gaussian kernel density estimate of the given data x has k modes.

Examples

h.crit(rnorm(10), k = 1)

jenzopr/silvermantest documentation built on Nov. 4, 2019, 2:39 p.m.