chibar: simulation of chi-bar pvalues

Description

Function to simulate weights and pvalues of a chi-bar distribution for tests of type A and type B (Silvapulle MJ and Sen PK (2005) Constrained statistical inference, Wiley, New Jersey) on inequality constraints on the parameters of hlp, mph models.

Usage

chibar(m, Z, ZF, d.fct = 0, h.fct = 0, 
test0 = 0, test1 = 0, repli = 0, derdt.fct = 0,
derht.fct = 0, formula = NULL, names = NULL, lev)

Arguments

Estimated expected frequencies

Population matrix - see the help of `mphineq.fit'

Sample matrix - see the help of `mphineq.fit'

d.fct

Inequality constraints function - see the help of `mphineq.fit'

h.fct

Equality constraints function - see the help of `mphineq.fit'

test0

Likelihood ratio statistics for testing problems of type A (Silvapulle and Sen, 2005)

test1

Likelihood ratio statistics for testing problems of type B (Silvapulle and Sen, 2005)

repli

Number of simulations

derdt.fct

Derivatives of inequality constraints - see the help of `mphineq.fit'

derht.fct

Derivatives of equality constraints - see the help of `mphineq.fit'

formula

Formula of the reference log-linear model (see details in `hmmm.model')

names

A character vector whose elements are the names of the variables

lev

Number of categories of the variables

Value

A list with the likelihood ratio statistics for testing hypotheses of type A and B (Silvapulle and Sen, 2005, pg. 61) and their simulated pvalues.

Details

The method "Simulation 2" described in Silvapulle and Sen, 2005, pg. 79 is used.

References

Silvapulle MJ, Sen PK (2005) Constrained statistical inference, Wiley, New Jersey.