Home > brede > brede_pde_binmix.m

brede_pde_binmix

PURPOSE ^

brede_pde_binmix - Estimation of binomial mixture model

SYNOPSIS ^

function O = brede_pde_binmix(NN, varargin)

DESCRIPTION ^

 brede_pde_binmix     - Estimation of binomial mixture model

       O = brede_pde_binmix(N)

       Input:    N      Observed frequencies

       Property: Info   Amount of information. A value over
                        5 gives information for each iteration

       Output:   O      Parameters [ P1 P2 lambda ]

       Estimation of mixture of two binomials (binomial mixture)

       The first element in N should contain the observed frequency
       zero positive outcomes, the second the observed frequency of
       one positive outcomes, the third the observed frequency of two
       positive outcomes, ...
 
       The input argument 'N' can also be a matrix. The rows are then
       modeled with a common lambda.

       The parameters in the output argument is ordered as: P1, P2,
       lambda, where lambda is the mixing coefficient. The return
       argument is sorted so that P1<=P2.

       Ref: John Uebersax, MIXBIN.
            Gelfand AE, Solomon A, JASA 69:32

       Example:
         N = [ 5 0 2 2 3 ];
         brede_pde_binmix(N)

         % Yerushalmy radiograph data from MIXBIN
         %   P1=0.0100  P2=0.6446, lambda= 0.9808
         N = [ 13560 877 168 66 42 28 23 39 64];
         brede_pde_binmix(N, 'info', 5)
 
       See also BREDE, BREDE_PDE_BINMIX_LL, BREDE_PDE_BINMIX_DER,
                BREDE_PDE_SPECHT, BREDE_OPT_CG, BREDE_OPT_GRADIENT.

 $Id: brede_pde_binmix.m,v 1.7 2007/09/21 16:19:57 fn Exp $

CROSS-REFERENCE INFORMATION ^

This function calls: This function is called by:
Generated on Fri 27-Nov-2009 18:11:22 by m2html © 2005