//// TD learning
clear; clf; 
gamma0=0.8; reward=[0 0 1 0 0]; pattern_vector=zeros(10,5);
pattern_vector(:,3)=[1;0;0;0;0;0;0;0;0;0];
w=zeros(1,10); V_mem=0; previous_state=5;
   
for episode=1:100; TDerror(episode)=0;
    if episode==50; pattern_vector(:,2)=[0;0;0;1;0;0;0;0;0;0]; end;
    for pattern=1:5
        V=w*pattern_vector(:,pattern);
        rhat(episode,pattern)=reward(previous_state)+gamma0*V-V_mem;
        w=w+0.2*rhat(episode,pattern)*pattern_vector(:,previous_state)';
        TDerror(episode)=TDerror(episode)+abs(rhat(episode,pattern));
        previous_state=pattern; V_mem=V;
    end
end

plot(rhat); xlabel('Episode'); ylabel('rhat');