Mercurial > hg > camir-aes2014
annotate toolboxes/FullBNT-1.0.7/HMM/pomdp_sample.m @ 0:e9a9cd732c1e tip
first hg version after svn
author | wolffd |
---|---|
date | Tue, 10 Feb 2015 15:05:51 +0000 |
parents | |
children |
rev | line source |
---|---|
wolffd@0 | 1 function [obs, hidden] = pomdp_sample(initial_prob, transmat, obsmat, act) |
wolffd@0 | 2 % SAMPLE_POMDP Generate a random sequence from a Partially Observed Markov Decision Process. |
wolffd@0 | 3 % [obs, hidden] = sample_pomdp(prior, transmat, obsmat, act) |
wolffd@0 | 4 % |
wolffd@0 | 5 % Inputs: |
wolffd@0 | 6 % prior(i) = Pr(Q(1)=i) |
wolffd@0 | 7 % transmat{a}(i,j) = Pr(Q(t)=j | Q(t-1)=i, A(t)=a) |
wolffd@0 | 8 % obsmat(i,k) = Pr(Y(t)=k | Q(t)=i) |
wolffd@0 | 9 % act(a) = A(t), so act(1) is ignored |
wolffd@0 | 10 % |
wolffd@0 | 11 % Output: |
wolffd@0 | 12 % obs and hidden are vectors of length T=length(act) |
wolffd@0 | 13 |
wolffd@0 | 14 |
wolffd@0 | 15 len = length(act); |
wolffd@0 | 16 hidden = mdp_sample(initial_prob, transmat, act); |
wolffd@0 | 17 obs = zeros(1, len); |
wolffd@0 | 18 for t=1:len |
wolffd@0 | 19 obs(t) = sample_discrete(obsmat(hidden(t),:)); |
wolffd@0 | 20 end |