1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
|
function [a, b, XLO, XUP] = find_invgampar(XLO, XUP, PLO, PUP, a0, b0);
% This function takes as inputs the bounds [XLO, XUP] in the support of the
% Inverse Gamma distribution (with parameters a and b), the probabilities of the
% bounds [PLO, PUP], and the initial values for ab=[a0, b0]
% and returns the estimates of a and b (as well as XLO and XUP)
% by solving the non-linear functions in invgampar(ab, XLO, XUP, PLO, PUP).\
%-----------------------------------------------------------------------------------
%------------------------ Inverse-Gamma distribution ------------------------------%
%--- p(x) = ( b^a/Gamma(a) ) x^(-a-1) exp(-b/x) for a>0 and b>0.
%--- where a is shape and b is scale parameter.
%--- E(x) = b/(a-1) for a>1; var(x) = b^2/( (a-1)^2*(a-2) ) for a>2;
%--- Noninformative distribution: a,b -> 0.
%--- How to draw: (1) draw z from Gamma(a,b); (2) let x=1/z.
%-----------------------------------------------------------------------------------
%
% Copyright (C) 1997-2012 Tao Zha
%
% This free software: you can redistribute it and/or modify
% it under the terms of the GNU General Public License as published by
% the Free Software Foundation, either version 3 of the License, or
% (at your option) any later version.
%
% It is distributed in the hope that it will be useful,
% but WITHOUT ANY WARRANTY; without even the implied warranty of
% MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
% GNU General Public License for more details.
%
% If you did not received a copy of the GNU General Public License
% with this software, see <http://www.gnu.org/licenses/>.
%
if XLO >= XUP;
error('the lower bound needs to be smaller than the upper bound')
elseif XLO<= 0;
error('the support for Gamma distribution needs to be non-negative');
end;
if a0 <= 0 || b0 <= 0;
error('the values for a and b need to be positive');
end;
disp(' ')
disp('*************** Convergence results for inverse gamma density ***************')
options = optimset('Display', 'on','TolFun', 1.0e-10, 'TolX', 1.0e-10);
ab_values = fsolve('invgampar', [a0, b0], options, XLO, XUP, PLO, PUP);
a = ab_values(1); b = ab_values(2);
% Alternatively, it is possible to constrain the search for the values of a
% and b in the positive range (with 0 being the explicit lower bound) by
% using the lsqnonlin function (see below) instead of the fsolve. The
% tradeoff is that lsqnonlin is typically slower than fsolve.
% LB_a = 0; LB_b = 0; UB_a = Inf; UB_b = Inf;
% ab_values = lsqnonlin('invgampar', [a0, b0], [LB_a, LB_b], [UB_a, UB_b],...
% options, XLO, XUP, PLO, PUP);
% a = ab_values(1); b = ab_values(2);
|