robustmeasure: Dispersion statistics for small samples (general)
In thiloklein/dispersion: Dispersion Statistics for Small Samples

Description Usage Arguments Value Author(s) References Examples

This program generates sampling distributions of CoV, Gini, Entropy(H), GE, HHI, and P(S more equal than s), P(S less equal than s) from draws from Pareto distribution, for various alpha values. Also generates data for the power curve for all the indices CoV, Gini, Entropy(H), GE, HHI and P(S more equal than s), P(S less equal than s) for the left end of the samp dist for chosen size of the test: ie. prop from the true dist to the left of the 5# (1#) cutoff from the null hypothesized distribution for the right end of the samp dist for chosen size of the test: ie. prop from the true dist to the right of the 95# (99#) cutoff from the null hypothesized dist.

1	robustmeasure(s, shape, l1, n, bss, lb, ub)

`shape`	Gamma distribution parameter for shape, for D(1), shape=1: uniform distribution on the simplex
`s`	The observed sample (not in share vector form).
`n`	bootstrap sample size , typically 3
`bss`	number of bootstrap samples
`lb`	lower bound 5 or .05 or 10 or .1
`ub`	upper bound 95 or .95 or 90 or .9
`l1`	the number of samples drawn for prior - S~D(lambda,) to calculate by simulation

robustmeasure returns samp dists of different measures for different alphas and power test proportions at the right of the right side cutoff and left of the left side cutoff of the samp dist for different measures.

Paul Kattuman, Thilo Klein, Jun Ma

Kattuman, P.A. and Ma, J. (2014). Dispersion Statistics for Small Samples. Working Paper, Cambridge Judge Business School.

## share vector of length 3
robustmeasure(shape=1,s=c(1,3,7),n=3,bss=2000,lb=0.1,ub=0.9,l1=1000)

## share vector of length larger 3
robustmeasure(shape=1,s=c(1,2,3,4,5,6,7,8,9,10),n=3,bss=2000,lb=0.1,ub=0.9,l1=100)

## Not run: 
## matrix of share vectors
data(libor)
robustmeasure(s=libor, shape=1, l1=100, n=3, bss=2000, lb=10, ub=90)

## End(Not run)