binom_betaPostPredDemo

Contents

Posterior Predictive Distribution for Beta-Binomial model
Create the model
Prior Predictive
Posterior Predictive
MAP estimation

Posterior Predictive Distribution for Beta-Binomial model

%#testPMTK
N = 10;
X = [1 2]';
N1 = sum(X);
N0 = N*length(X)-N1;
a = 2;
b = 2;

Create the model

m = Binom_BetaDist('N',N,'prior', BetaDist(a,b));

Prior Predictive

m0 = marginal(m);                   % BetaBinommDist object
figure; plot(m0); title('prior predictive')

Posterior Predictive

m = fit(m, 'data', X);
mm = marginal(m);                   % posterior predictive is BetaBinomDist
figure;
h1 = plot(mm, 'plotArgs', 'b'); title('posterior predictive')

MAP estimation

m3 = fit(BinomDist('N',N), 'data', X, 'prior', BetaDist(a,b));
hold on
h3 = plot(m3);
set(h3, 'faceColor','r','barwidth',0.5);
legend('postpred', 'plugin');