gbmodes1

Usage

Overview

Multimodality analysis via kernel density estimation. Compute the maximal kernel bandwidth h_c(M) for which #(modes in the estimated density)>M.The significance of h_c is computed via smoothed bootstrap. Read data from std. input. Print the couple h_c p(h_c) and the modal values (set with -O).

Options

-n

number of equispaced points where the density is computed (default 100)

-m

number of modes (default 1)

-r

search range for modes; comma separated couple 'data_min,data_max'

-s

scale the search range to [min*s+(1-s)*Max,Max*s+(1-s)*min] (default 1)

-e

relative tolerance on h_c value (default 1e-6)

-S

significance level (for Hall-York correction) (default .05)

-t

number of bootstrap trials used for significance computation (default 1000)

-R

RNG seed for bootstrap trials (default 0)

-K

choose the kernel to use: 0 Gaussian or 1 Laplacian (default 0)

-v

verbose mode (more verbose if provided 2 times)

-O

output type (default 0)

-F

specify the input fields separators (default " \t")

-h

this help