% straight forward outline of running of scripts

%This used the first count observations and predicts expected & the
%standard deviation of obs(count+1). 
    display('outline')
    display('basic parameters')
basics_f
p(1:3)=[.09 .01 .05];
r= p(2) + p(1)*exp(-p(3).*(0:m+m2)); %r= failure rate
    rel = exp(-r); %reliabilty of a component 
    pf = 1-rel; %prob component fails before 1 month
    display('simulating fails per month')
fails_per_month_f
    display('sampling from the prior distribution')
    
    obs(1:m+m2) = K(1,1:m+m2); %this gives I observed #failures in months 1:I. Only one obs for each month. 
                     % note that the first obs is 0 since 0 fails in first
                      % month.    
    sampling_f
for count=no_pred:I   %use the first "count" obs to predict the next one.                     
count

    display('samping from the posterior using importance sampling')
post_sampling_f
    display('sort and display results')
sort_disp
    display('posterior predictive')

    post_pred_correct

end
%upper(1:I+1) =post_prediction(1:(I+1)) + 2*sqrt(post_var(1:(I+1)));
%lower(1:I+1) =post_prediction(1:(I+1)) - 2*sqrt(post_var(1:(I+1)));

figure(111)
hold off
plot(2:I,obs(2:I),'b.-')
hold on
plot((no_pred+1):(I+1),post_prediction((no_pred+1):(I+1)),'g.-')
plot((no_pred+1):(I+1),upper((no_pred+1):(I+1)),'r.:')
plot((no_pred+1):(I+1),lower((no_pred+1):(I+1)),'m.:')
plot(I:(I+1),obs(I:(I+1)),'k.:')
axis([0 I+2 0 (1.2*n)])
hold off