steps = 20;
% steps = 5;
dims = 5;
d_steps = 100000;      % precision in defining the distributions of the conditional probability of failure (in partitions)
% char ind_space;
inds = Lin_Mat(dims);
for i=1:length(inds),
    inds(i) = 1;
end
% alpha = [10 50 20 50 20];
alpha = [10 10 40 30 10];       % Bootstrap Dirichlet distribution

% Date: 17/06/2023

% Now calculate the marginal distribution of the total probability of failure as a sum of convolutions.
% need to define the distributions of the conditional probability of
% failure in sub-domains.
% 
% Bootstraping distribution (prior for all AVs)
cond_distr_params = struct('partition',[],'alfa',[],'beta',[],'steps',[],'pdf',[]);
cond_distr_params(1).partition = 1;
cond_distr_params(1).alfa = 2;
cond_distr_params(1).beta = 299;
% cond_distr_params(1).steps = 10000;
cond_distr_params(1).steps = d_steps;
cond_distr_params(2).partition = 2;
cond_distr_params(2).alfa = 2;
cond_distr_params(2).beta = 800;
% cond_distr_params(2).steps = 10000;
cond_distr_params(2).steps = d_steps;
cond_distr_params(3).partition = 3;
cond_distr_params(3).alfa = 2;
cond_distr_params(3).beta = 1500;
% cond_distr_params(3).steps = 10000;
cond_distr_params(3).steps = d_steps;
cond_distr_params(4).partition = 4;
cond_distr_params(4).alfa = 2;
cond_distr_params(4).beta = 1000;
% cond_distr_params(4).steps = 10000;
cond_distr_params(4).steps = d_steps;
cond_distr_params(5).partition = 5;
cond_distr_params(5).alfa = 1;
cond_distr_params(5).beta = 400;
% cond_distr_params(5).steps = 10000;
cond_distr_params(5).steps = d_steps;
for i=1:dims,
    % get the distrbutions of the conditional probabilities of failure
    cond_distr_params(i).pdf = gen_beta_m(cond_distr_params(i).alfa,cond_distr_params(i).beta,cond_distr_params(i).steps);
end
% Observations for all AVs
AV_count = 5;
observ_AV1 = struct('partition',[],'miles',[],'failures',[]);
% observation AV1
observ_AV1(1).partition = 1;
observ_AV1(1).miles = 7;
observ_AV1(1).failures = 0;
observ_AV1(2).partition = 2;
observ_AV1(2).miles = 9;
observ_AV1(2).failures = 0;
observ_AV1(3).partition = 3;
observ_AV1(3).miles = 45;
observ_AV1(3).failures = 0;
observ_AV1(4).partition = 4;
observ_AV1(4).miles = 30;
observ_AV1(4).failures = 0;
observ_AV1(5).partition = 5;
observ_AV1(5).miles = 9;
observ_AV1(5).failures = 0;
total_miles(1) = 0;
total_failures(1) = 0; 
for i=1:AV_count,
    total_miles(1) = total_miles(1) + observ_AV1(i).miles;          % Count of miles used in black-box inference
    total_failures(1) = total_failures(1) + observ_AV1(i).failures; % Count of failures used in black-box inference
end

% observation AV2
observ_AV2 = struct('partition',[],'miles',[],'failures',[]);

observ_AV2(1).partition = 1;
observ_AV2(1).miles = 10;
observ_AV2(1).failures = 0;
observ_AV2(2).partition = 2;
observ_AV2(2).miles = 45;
observ_AV2(2).failures = 0;
observ_AV2(3).partition = 3;
observ_AV2(3).miles = 30;
observ_AV2(3).failures = 0;
observ_AV2(4).partition = 4;
observ_AV2(4).miles = 8;
observ_AV2(4).failures = 0;
observ_AV2(5).partition = 5;
observ_AV2(5).miles = 7;
observ_AV2(5).failures = 0;
total_miles(2)= 0;
total_failures(2) = 0; 
for i=1:AV_count,
    total_miles(2) = total_miles(2) + observ_AV2(i).miles;          % Count of miles used in black-box inference
    total_failures(2) = total_failures(2) + observ_AV2(i).failures; % Count of failures used in black-box inference
end

% observation AV3
observ_AV3 = struct('partition',[],'miles',[],'failures',[]);

observ_AV3(1).partition = 1;
observ_AV3(1).miles = 45;
observ_AV3(1).failures = 0;
observ_AV3(2).partition = 2;
observ_AV3(2).miles = 30;
observ_AV3(2).failures = 0;
observ_AV3(3).partition = 3;
observ_AV3(3).miles = 7;
observ_AV3(3).failures = 0;
observ_AV3(4).partition = 4;
observ_AV3(4).miles = 9;
observ_AV3(4).failures = 0;
observ_AV3(5).partition = 5;
observ_AV3(5).miles = 9;
observ_AV3(5).failures = 0;
total_miles(3) = 0;
total_failures(3) = 0; 
for i=1:AV_count,
    total_miles(3) = total_miles(3) + observ_AV3(i).miles;          % Count of miles used in black-box inference
    total_failures(3) = total_failures(3) + observ_AV3(i).failures; % Count of failures used in black-box inference
end

% observation AV3_fail
observ_AV3_fail = struct('partition',[],'miles',[],'failures',[]);

observ_AV3_fail(1).partition = 1;
observ_AV3_fail(1).miles = 45;
observ_AV3_fail(1).failures = 1;
observ_AV3_fail(2).partition = 2;
observ_AV3_fail(2).miles = 30;
observ_AV3_fail(2).failures = 1;
observ_AV3_fail(3).partition = 3;
observ_AV3_fail(3).miles = 7;
observ_AV3_fail(3).failures = 0;
observ_AV3_fail(4).partition = 4;
observ_AV3_fail(4).miles = 9;
observ_AV3_fail(4).failures = 0;
observ_AV3_fail(5).partition = 5;
observ_AV3_fail(5).miles = 9;
observ_AV3_fail(5).failures = 0;
total_miles(6) = 0;
total_failures(6) = 0; 
for i=1:AV_count,
    total_miles(6) = total_miles(6) + observ_AV3_fail(i).miles;          % Count of miles used in black-box inference
    total_failures(6) = total_failures(6) + observ_AV3_fail(i).failures; % Count of failures used in black-box inference
end

% observation AV4
observ_AV4 = struct('partition',[],'miles',[],'failures',[]);

observ_AV4(1).partition = 1;
observ_AV4(1).miles = 20;
observ_AV4(1).failures = 0;
observ_AV4(2).partition = 2;
observ_AV4(2).miles = 20;
observ_AV4(2).failures = 0;
observ_AV4(3).partition = 3;
observ_AV4(3).miles = 20;
observ_AV4(3).failures = 0;
observ_AV4(4).partition = 4;
observ_AV4(4).miles = 20;
observ_AV4(4).failures = 0;
observ_AV4(5).partition = 5;
observ_AV4(5).miles =20;
observ_AV4(5).failures = 0;
total_miles(4) = 0;
total_failures(4) = 0; 
for i=1:AV_count,
    total_miles(4) = total_miles(4) + observ_AV4(i).miles;          % Count of miles used in black-box inference
    total_failures(4) = total_failures(4) + observ_AV4(i).failures; % Count of failures used in black-box inference
end

% observation AV5
observ_AV5 = struct('partition',[],'miles',[],'failures',[]);
observ_AV5(1).partition = 1;
observ_AV5(1).miles = 45;
observ_AV5(1).failures = 0;
observ_AV5(2).partition = 2;
observ_AV5(2).miles = 19;
observ_AV5(2).failures = 0;
observ_AV5(3).partition = 3;
observ_AV5(3).miles = 7;
observ_AV5(3).failures = 0;
observ_AV5(4).partition = 4;
observ_AV5(4).miles = 9;
observ_AV5(4).failures = 0;
observ_AV5(5).partition = 5;
observ_AV5(5).miles = 20;
observ_AV5(5).failures = 0;
total_miles(5) = 0;
total_failures(5) = 0; 
% compute the counts needed for black-box inference
for i=1:AV_count,
    total_miles(5) = total_miles(5) + observ_AV5(i).miles;          % Count of miles used in black-box inference
    total_failures(5) = total_failures(5) + observ_AV5(i).failures; % Count of failures used in black-box inference
end

% observation All AVs
observ_all = struct('partition',[],'miles',[],'failures',[]);
observ_all(1).partition = 1;
observ_all(1).miles = 0;
observ_all(1).failures = 0;
observ_all(2).partition = 2;
observ_all(2).miles = 0;
observ_all(2).failures = 0;
observ_all(3).partition = 3;
observ_all(3).miles = 0;
observ_all(3).failures = 0;
observ_all(4).partition = 4;
observ_all(4).miles = 0;
observ_all(4).failures = 0;
observ_all(5).partition = 5;
observ_all(5).miles = 0;
observ_all(5).failures = 0;
vendor_total_miles = 0;     % Count of miles used for vendor black-box inference
vendor_total_failures = 0;  % Count of failures used for vendor black-box inference


% Compute the aggregated counts (miles and failures for all AVs)
for i=1:AV_count,  
  observ_all(i).miles = observ_all(i).miles + observ_AV1(i).miles;
  observ_all(i).miles = observ_all(i).miles + observ_AV2(i).miles;
  observ_all(i).miles = observ_all(i).miles + observ_AV3(i).miles;
  observ_all(i).miles = observ_all(i).miles + observ_AV4(i).miles;
  observ_all(i).miles = observ_all(i).miles + observ_AV5(i).miles;
  vendor_total_miles = vendor_total_miles + observ_all(i).miles;
  observ_all(i).failures = observ_all(i).failures + observ_AV1(i).failures;
  observ_all(i).failures = observ_all(i).failures + observ_AV2(i).failures;
  observ_all(i).failures = observ_all(i).failures + observ_AV3(i).failures;
  observ_all(i).failures = observ_all(i).failures + observ_AV4(i).failures;
  observ_all(i).failures = observ_all(i).failures + observ_AV5(i).failures;
  vendor_total_failures = vendor_total_failures + observ_all(i).failures;
end

% observation All AVs with failures
observ_all_fail = struct('partition',[],'miles',[],'failures',[]);
observ_all_fail(1).partition = 1;
observ_all_fail(1).miles = 0;
observ_all_fail(1).failures = 0;
observ_all_fail(2).partition = 2;
observ_all_fail(2).miles = 0;
observ_all_fail(2).failures = 0;
observ_all_fail(3).partition = 3;
observ_all_fail(3).miles = 0;
observ_all_fail(3).failures = 0;
observ_all_fail(4).partition = 4;
observ_all_fail(4).miles = 0;
observ_all_fail(4).failures = 0;
observ_all_fail(5).partition = 5;
observ_all_fail(5).miles = 0;
observ_all_fail(5).failures = 0;
vendor_total_miles_fail = 0;     % Count of miles used for vendor black-box inference
vendor_total_failures_fail = 0;  % Count of failures used for vendor black-box inference

% Compute the aggregatesd counts (miles and failures for all AVs WITH FAILURES)
for i=1:AV_count,  
  observ_all_fail(i).miles = observ_all(i).miles + observ_AV1(i).miles;
  observ_all_fail(i).miles = observ_all(i).miles + observ_AV2(i).miles;
  observ_all_fail(i).miles = observ_all(i).miles + observ_AV3_fail(i).miles;
  observ_all_fail(i).miles = observ_all(i).miles + observ_AV4(i).miles;
  observ_all_fail(i).miles = observ_all(i).miles + observ_AV5(i).miles;
  vendor_total_miles_fail = vendor_total_miles_fail + observ_all(i).miles;
  observ_all_fail(i).failures = observ_all(i).failures + observ_AV1(i).failures;
  observ_all_fail(i).failures = observ_all(i).failures + observ_AV2(i).failures;
  observ_all_fail(i).failures = observ_all(i).failures + observ_AV3_fail(i).failures;
  observ_all_fail(i).failures = observ_all(i).failures + observ_AV4(i).failures;
  observ_all_fail(i).failures = observ_all(i).failures + observ_AV5(i).failures;
  vendor_total_failures_fail = vendor_total_failures_fail + observ_all(i).failures;
end

% Define a set of profiles
profile = struct('P_part_1',[],'P_part_2',[],'P_part_3',[],'P_part_4',[],'P_part_5',[],'name',[]);
profile(1).P_part_1 = 0.1;
profile(1).P_part_2 = 0.2;
profile(1).P_part_3 = 0.4;
profile(1).P_part_4 = 0.25;
profile(1).P_part_5 = 0.05;
profile(1).name = 'Operational profile 1';
profile(2).P_part_1 = 0.4;
profile(2).P_part_2 = 0.2;
profile(2).P_part_3 = 0.1;
profile(2).P_part_4 = 0.25;
profile(2).P_part_5 = 0.05;
profile(2).name = 'Operational profile 2';
profile(3).P_part_1 = 0.2;
profile(3).P_part_2 = 0.2;
profile(3).P_part_3 = 0.2;
profile(3).P_part_4 = 0.2;
profile(3).P_part_5 = 0.2;
profile(3).name = 'Operational profile 3';

% Define a set of structures for the results of all computations 

plot_array_descr = struct('Plot',[]);
% figure_n,title_text,X,Y,LD 
plot_descr = struct('X',[],'XLabel',[],'Y',[],'YLabel',[],'LD',[]);
plot_array_descr = struct('Plot',[]);
plot_descr.X = double([1:cond_distr_params(1).steps]/cond_distr_params(1).steps);
plot_descr.XLabel = 'pfm';
plot_descr.Y = cond_distr_params(1).pdf
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OCl prior PFM'; 
%
plot_array_descr(1).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(2).steps]/cond_distr_params(2).steps);
plot_descr.XLabel = 'pfm';
plot_descr.Y = cond_distr_params(2).pdf;
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC2 prior PFM'; 
plot_array_descr(2).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(3).steps]/cond_distr_params(3).steps);
plot_descr.XLabel = 'pfm';
plot_descr.Y = cond_distr_params(3).pdf;
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC3 prior PFM'; 
plot_array_descr(3).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(4).steps]/cond_distr_params(4).steps);
plot_descr.XLabel = 'pfm';
plot_descr.Y = cond_distr_params(4).pdf;
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC4 prior PFM'; 
plot_array_descr(4).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'pfm';
plot_descr.Y = cond_distr_params(5).pdf;
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC5 prior PFM'; 
plot_array_descr(5).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Marginal PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Operational profile 1'; 
plot_array_descr(6).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Marginal PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Operational profile 2'; 
plot_array_descr(7).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Marginal PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Operational profile 3'; 
plot_array_descr(8).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'conditional PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC1: posterior PFM, Observation 1'; 
plot_array_descr(9).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'conditional PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC2: posterior PFM, Observation 1'; 
plot_array_descr(10).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'conditional PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC3: posterior PFM, Observation 1'; 
plot_array_descr(11).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'conditional PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC4: posterior PFM, Observation 1'; 
plot_array_descr(12).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'conditional PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC5: posterior PFM, Observation 1'; 
plot_array_descr(13).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'conditional PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC1: posterior PFM, Observation 2'; 
plot_array_descr(14).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'conditional PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC2: posterior PFM, Observation 2'; 
plot_array_descr(15).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'conditional PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC3: posterior PFM, Observation 2'; 
plot_array_descr(16).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'conditional PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC4: posterior PFM, Observation 2'; 
plot_array_descr(17).Plot = plot_descr;
%
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'conditional PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'OC5: posterior PFM, Observation 2'; 
plot_array_descr(18).Plot = plot_descr;

% Bootstrap place holder
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Marginal prior PFM (bootstrap)'; 
plot_array_descr(19).Plot = plot_descr;

% AV 1 placeholders
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV1: Marginal prior PFM'; 
plot_array_descr(20).Plot = plot_descr;

% AV1 black-box posterior distribution placeholder
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV1: BB marginal posterior PFM (own data)'; 
plot_array_descr(21).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV1: BB marginal posterior PFM (fleet data)'; 
plot_array_descr(22).Plot = plot_descr;

% AV1 white-box inference
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV1: WB marginal posterior PFM (own data)'; 
plot_array_descr(23).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV1: WB marginal posterior PFM(fleet data)'; 
plot_array_descr(24).Plot = plot_descr;
%
% AV 2 placeholders
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV2: Marginal prior PFM'; 
plot_array_descr(25).Plot = plot_descr;

% AV2 black-box posterior distribution placeh
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV2: BB marginal posterior PFM(own data)'; 
plot_array_descr(26).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV2: BB marginal posterior PFM (fleet data)'; 
plot_array_descr(27).Plot = plot_descr;

% AV2 white-box inference
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV2: WB marginal posterior PFM (own data)'; 
plot_array_descr(28).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Marginal system PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV2: WB marginal posterior PFM (fleet data)'; 
plot_array_descr(29).Plot = plot_descr;

% AV 3 placeholders
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: Marginal prior PFM'; 
plot_array_descr(30).Plot = plot_descr;

% AV3 black-box posterior distribution placeh
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: BB marginal posterior PFM (own data)'; 
plot_array_descr(31).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: BB marginal posterior PFM (fleet data)'; 
plot_array_descr(32).Plot = plot_descr;

% AV3 white-box inference
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: WB marginal posterior PFM (own data)'; 
plot_array_descr(33).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: WB marginal posterior PFM (all fleet)'; 
plot_array_descr(34).Plot = plot_descr;

% AV 4 placeholders
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV4: Marginal prior PFM'; 
plot_array_descr(35).Plot = plot_descr;

% AV4 black-box posterior distribution placeh
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV4: BB marginal posterior PFM (own data)'; 
plot_array_descr(36).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV4: BB marginal posterior PFM (fleet data)'; 
plot_array_descr(37).Plot = plot_descr;

% AV4 white-box inference
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV4: WB marginal posterior PFM (own data)'; 
plot_array_descr(38).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV4: WB marginal posterior PFM (fleet data)'; 
plot_array_descr(39).Plot = plot_descr;

% AV 5 placeholders
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV5: Marginal prior PFM'; 
plot_array_descr(40).Plot = plot_descr;

% AV5 black-box posterior distribution placeh
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV5: BB marginal posterior PFM (own data)'; 
plot_array_descr(41).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV5: BB marginal posterior PFM (fleet data)'; 
plot_array_descr(42).Plot = plot_descr;

% AV5 white-box inference
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV5: WB marginal posterior PFM (own data)'; 
plot_array_descr(43).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV5: WB marginal posterior PFM (fleet data)'; 
plot_array_descr(44).Plot = plot_descr;

% Vendor placeholders
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: Marginal prior PFM'; 
plot_array_descr(45).Plot = plot_descr;

% Black-box placeholders
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: BB marginal posterior PFM (fleet data)'; 
plot_array_descr(46).Plot = plot_descr;

% White-box placeholders
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: WB marginal posterior PFM (fleet data)'; 
plot_array_descr(47).Plot = plot_descr;

% Observations with failures

% AV1 
% Black - box placeholder (fleet data)
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV1: BB marginal posterior PFM (fleet data with failures)'; 
plot_array_descr(48).Plot = plot_descr;

% White-box placeholder (fleet data)
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV1: WB marginal posterior PFM (fleet data with failures)'; 
plot_array_descr(49).Plot = plot_descr;

% AV2 
% Black - box placeholder (fleet data)
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV2: BB marginal posterior PFM (fleet data with failures)'; 
plot_array_descr(50).Plot = plot_descr;

% White-box placeholder (fleet data)
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV2: WB marginal posterior PFM (fleet data with failures)'; 
plot_array_descr(51).Plot = plot_descr;

% AV3
% Black-box placeholders
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: BB marginal posterior PFM (own data with failures)'; 
plot_array_descr(52).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: BB marginal Posterior PFM (fleet data with failures)'; 
plot_array_descr(53).Plot = plot_descr;

% White-box placeholders
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: WB marginal posterior PFM (own data with failures)'; 
plot_array_descr(54).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: WB marginal posterior PFM (fleet data with failures)'; 
plot_array_descr(55).Plot = plot_descr;

% AV4
% Black - box placeholder (fleet data)
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV4: BB marginal posterior PFM (fleet data with failures)'; 
plot_array_descr(56).Plot = plot_descr;

% White-box placeholder (fleet data)
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'WB marginal posterior PFM (AV4: fleet data with failures)'; 
plot_array_descr(57).Plot = plot_descr;

% AV5 
% Black - box placeholder (fleet data)
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV5: BB marginal posterior PFM (fleet data with failures)'; 
plot_array_descr(58).Plot = plot_descr;

% White-box placeholder (fleet data)
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV5: WB marginal posterior PFM (fleet data with failures)'; 
plot_array_descr(59).Plot = plot_descr;

% Vendor with failures 
% Black - box placeholder (with failures)
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: BB marginal posterior PFM (fleet data with failures)'; 
plot_array_descr(60).Plot = plot_descr;

% White-box placeholder (with failures)
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'System PFM';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: WB marginal posterior PFM (fleet data with failures)'; 
plot_array_descr(61).Plot = plot_descr;

% Posterior conditional distributions
% No failures observed

% AV1 
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC1)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV1: posterior PFM (OC1: own data)'; 
plot_array_descr(62).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC2)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV1: posterior PFM (OC2: own data)'; 
plot_array_descr(63).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC3)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV1: posterior PFM (OC3: own data)'; 
plot_array_descr(64).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC4)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV1: posterior PFM (OC4: own data)'; 
plot_array_descr(65).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC5)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV1: posterior PFM (OC5: own data)'; 
plot_array_descr(66).Plot = plot_descr;

% AV2
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC1)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV2: posterior PFM (OC1: own data)'; 
plot_array_descr(67).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC2)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV2: posterior PFM (OC2: own data)'; 
plot_array_descr(68).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC3)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV2: posterior PFM (OC3: own data)'; 
plot_array_descr(69).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC4)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV2: posterior PFM (OC4: own data)'; 
plot_array_descr(70).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC5)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV2: posterior PFM (OC5: own data)'; 
plot_array_descr(71).Plot = plot_descr;

% AV3
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC1)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: posterior PFM (OC1: own data)'; 
plot_array_descr(72).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC2)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: posterior PFM (OC2: own data)'; 
plot_array_descr(73).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC3)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: posterior PFM (OC3: own data)'; 
plot_array_descr(74).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC4)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: posterior PFM (OC4: own data)'; 
plot_array_descr(75).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC5)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: posterior PFM (OC5: own data)'; 
plot_array_descr(76).Plot = plot_descr;

% AV4 
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC1)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV4: posterior PFM (OC1: own data)'; 
plot_array_descr(77).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC2)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV4: posterior PFM (OC2: own data)'; 
plot_array_descr(78).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC3)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV4: posterior PFM (OC3: own data)'; 
plot_array_descr(79).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC4)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV4: posterior PFM (OC4: own data)'; 
plot_array_descr(80).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC5)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV4: posterior PFM (OC5: own data)'; 
plot_array_descr(81).Plot = plot_descr;

% AV5 
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC1)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV5: posterior PFM (OC1: own data)'; 
plot_array_descr(82).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC2)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV5: posterior PFM (OC2: own data)'; 
plot_array_descr(83).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC3)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV5: posterior PFM (OC3: own data)'; 
plot_array_descr(84).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC4)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV5: posterior PFM (OC4: own data)'; 
plot_array_descr(85).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC5)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV5: posterior PFM (OC5: own data)'; 
plot_array_descr(86).Plot = plot_descr;

% Vendor 
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC1)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: posterior PFM (OC1: fleet data)'; 
plot_array_descr(87).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC2)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: posterior PFM (OC2: fleet data)'; 
plot_array_descr(88).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC3)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: posterior PFM (OC3: fleet data)'; 
plot_array_descr(89).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC4)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: posterior PFM (OC4: fleet data)'; 
plot_array_descr(90).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC5)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: posterior PFM (OC5: fleet data)'; 
plot_array_descr(91).Plot = plot_descr;

% Failures observed (by AV3)

% AV3 (own data)
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC1)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: posterior PFM (OC1: own data with failures)'; 
plot_array_descr(92).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC2)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: posterior PFM (OC2: own data with failures)'; 
plot_array_descr(93).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC3)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: posterior PFM (OC3: own data with failures)'; 
plot_array_descr(94).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC4)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: posterior PFM (OC4: own data with failures)'; 
plot_array_descr(95).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC5)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'AV3: posterior PFM (OC5: own data with failures)'; 
plot_array_descr(96).Plot = plot_descr;

% Vendor 
plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC1)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: posterior PFM (OC1: fleet data with failures)'; 
plot_array_descr(97).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC2)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: posterior PFM (OC2: fleet data with failures)'; 
plot_array_descr(98).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC3)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: posterior PFM (OC3: fleet data with failures)'; 
plot_array_descr(99).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC4)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: posterior PFM (OC4: fleet data with failures)'; 
plot_array_descr(100).Plot = plot_descr;

plot_descr.X = double([1:cond_distr_params(5).steps]/cond_distr_params(5).steps);
plot_descr.XLabel = 'Conditional PFM (OC5)';
plot_descr.Y = Lin_Mat(cond_distr_params(5).steps);
plot_descr.YLabel = 'probability density function';
plot_descr.LD = 'Vendor: posterior PFM (OC5: fleet data with failures)'; 
plot_array_descr(101).Plot = plot_descr;

marginal_distr = Lin_Mat(cond_distr_params(1).steps);
for i=1:3,
  mg = marginal_distr;
    for j = 1:5
      switch j
        case 1
           prob = profile(i).P_part_1;
        case 2
           prob = profile(i).P_part_2;
        case 3
           prob = profile(i).P_part_3;
        case 4
           prob = profile(i).P_part_4;
        otherwise
           prob = profile(i).P_part_5;
      end
      cond_distr = gen_beta_scaled(cond_distr_params(j).alfa,cond_distr_params(j).beta,prob,cond_distr_params(j).steps);
      if(j==1)
         marg_conv = cond_distr;
      else
         marg_conv = conv(marg_conv,cond_distr);
         marg_conv(length(marg_conv)+1) = double(0);
      end
    end    
    % now apply the waigth to this convolution
    mg = marginal_distr + marg_conv;
    plot_array_descr(i+5).Plot.Y = mg;
end
% Now plot all graphs in a single plot figure_n,title_text,X,Y,LD
P1(1) = plot_array_descr(2);
P1(2) = plot_array_descr(3);
P1(3) = plot_array_descr(5);
P2(1) = plot_array_descr(1);
P2(2) = plot_array_descr(4);
Data = struct('Plot',[],'xlabel',[],'ylabel',[],'xlim',[],'title',[]);
Data(1).Plot = P1;
Data(1).xlabel = P1(1).Plot.XLabel;
Data(1).ylabel = P1(1).Plot.YLabel;
Data(1).xlim = 0.03;
Data(1).title = 'Conditional Probability of failure (in OCs)';
Data(2).Plot = P2;
Data(2).xlabel = P2(1).Plot.XLabel;
Data(2).ylabel = P2(1).Plot.YLabel;
Data(2).xlim = 0.03;
Data(2).title = 'Conditional Probability of failure (in OCs)';
[fig_num,subplot_num] = plotting_subplots(1,Data,1,2,2);
clear Data;
% Now plot the profiles 
[fig_num,subplot_num] = plot_bar_subplots(1,profile,3,2,2);
% Now plot the marginal distribution of PFM for all profiles. 
P3(1) = plot_array_descr(6);
P3(2) = plot_array_descr(7);
P3(3) = plot_array_descr(8);
Data(1).Plot = P3;
Data(1).xlabel = P3(1).Plot.XLabel;
Data(1).ylabel = P3(1).Plot.YLabel;
Data(1).xlim = 0.03;
Data(1).title = 'Marginal Probability of failure';
[fig_num,subplot_num] = plotting_subplots(1,Data,4,2,2);
inds(dims) = 0;
count = 0;
ind_space = struct('indeces',[],'pdf',[],'volume',[],'x',[]);
while (steps > 0)   % infinite loop 
    [inds, done] = get_next_index(inds, steps);
    if(done > 0)
        if(sum(inds) == steps)
          count = count + 1;
          display(count);
          ind_space(count).indeces = inds;
        end
    end
    if(done > 1)
        break;
    end
end
display(count);
display(ind_space);
% Now convert the string into a cell of strings
x = Lin_Mat(dims);
integral = double(0);
vol_unit = power(double(1)/double(steps),dims-1);
for i=1:count,
    inds = ind_space(i).indeces;
    x = double(inds) / double(steps);
    display(x);
    ind_space(i).x = x;
    D = dirichlet_robust(x,alpha);
    ind_space(i).pdf = D;
    display(D);
    ind_space(i).volume = D * vol_unit;
    integral = integral + D * vol_unit;
end
display(integral);
integral2 = double(0);
for i=1:count,
 ind_space(i).volume = ind_space(i).volume / integral;   % compensate rounding errors and set the integral to 1. 
 integral2 = integral2 + ind_space(i).volume;
end

marginal_distr = Lin_Mat(cond_distr_params(1).steps);
for i=1:count,
    for j=1:dims,
        cond_distr = gen_beta_scaled(cond_distr_params(j).alfa,cond_distr_params(j).beta,ind_space(i).x(j),cond_distr_params(j).steps);
        if(j==1)
            marg_conv = cond_distr;
        else
            marg_conv = conv(marg_conv,cond_distr);
            marg_conv(length(marg_conv)+1) = double(0);
        end
    end
    % now apply the waigth to this convolution
    marg_conv = marg_conv * ind_space(i).volume; 
    % add add it to the marginal distribution
    marginal_distr = marginal_distr + marg_conv;
end
display(sum(marginal_distr));
% figure(100);
observ = struct('partition',[],'tests',[],'failures',[]);
% observation 1 - obsolete. Observations defined earlier in the code. 
% observ(1).partition = 1;
% observ(1).tests = 8;
% observ(1).failures = 0;
% observ(2).partition = 2;
% observ(2).tests = 22;
% observ(2).failures = 0;
% observ(3).partition = 3;
% observ(3).tests = 39;
% observ(3).failures = 0;
% observ(4).partition = 4;
% observ(4).tests = 23;
% observ(4).failures = 0;
% observ(5).partition = 5;
% observ(5).tests = 8;
% observ(5).failures = 0;

% AV1 - Update Beta parameters
cond_distr_params(6).partition = 1;
cond_distr_params(6).alfa = cond_distr_params(1).alfa + observ_AV1(1).failures;
cond_distr_params(6).beta = cond_distr_params(1).beta + observ_AV1(1).miles - observ_AV1(1).failures;
cond_distr_params(6).steps = cond_distr_params(1).steps;
cond_distr_params(7).partition = 2;
cond_distr_params(7).alfa = cond_distr_params(2).alfa + observ_AV1(2).failures;
cond_distr_params(7).beta = cond_distr_params(2).beta + observ_AV1(2).miles - observ_AV2(2).failures;
cond_distr_params(7).steps = cond_distr_params(2).steps;
cond_distr_params(8).partition = 3;
cond_distr_params(8).alfa = cond_distr_params(3).alfa + observ_AV1(3).failures;
cond_distr_params(8).beta = cond_distr_params(3).beta + observ_AV1(3).miles - observ_AV1(3).failures;
cond_distr_params(8).steps = cond_distr_params(3).steps;
cond_distr_params(9).partition = 4;
cond_distr_params(9).alfa = cond_distr_params(4).alfa + observ_AV1(4).failures;
cond_distr_params(9).beta = cond_distr_params(4).beta + observ_AV1(4).miles - observ_AV1(4).failures;
cond_distr_params(9).steps = cond_distr_params(4).steps;
cond_distr_params(10).partition = 5;
cond_distr_params(10).alfa = cond_distr_params(5).alfa + observ_AV1(5).failures;
cond_distr_params(10).beta = cond_distr_params(5).beta + observ_AV1(5).miles - observ_AV1(5).failures;
cond_distr_params(10).steps = cond_distr_params(5).steps;
alpha_obs_AV1 = alpha;

for i=1:length(alpha),
   alpha_obs_AV1(i) = alpha_obs_AV1(i) + observ_AV1(i).miles;
end
alpha_obs_AV1_fail = alpha_obs_AV1;


% AV2 - Update Beta parameters
cond_distr_params(11).partition = 1;
cond_distr_params(11).alfa = cond_distr_params(1).alfa + observ_AV2(1).failures;
cond_distr_params(11).beta = cond_distr_params(1).beta + observ_AV2(1).miles - observ_AV2(1).failures;
cond_distr_params(11).steps = cond_distr_params(1).steps;
cond_distr_params(12).partition = 2;
cond_distr_params(12).alfa = cond_distr_params(2).alfa + observ_AV2(2).failures;
cond_distr_params(12).beta = cond_distr_params(2).beta + observ_AV2(2).miles - observ_AV2(2).failures;
cond_distr_params(12).steps = cond_distr_params(2).steps;
cond_distr_params(13).partition = 3;
cond_distr_params(13).alfa = cond_distr_params(3).alfa + observ_AV2(3).failures;
cond_distr_params(13).beta = cond_distr_params(3).beta + observ_AV2(3).miles - observ_AV2(3).failures;
cond_distr_params(13).steps = cond_distr_params(3).steps;
cond_distr_params(14).partition = 4;
cond_distr_params(14).alfa = cond_distr_params(4).alfa + observ_AV2(4).failures;
cond_distr_params(14).beta = cond_distr_params(4).beta + observ_AV2(4).miles - observ_AV2(4).failures;
cond_distr_params(14).steps = cond_distr_params(4).steps;
cond_distr_params(15).partition = 5;
cond_distr_params(15).alfa = cond_distr_params(5).alfa + observ_AV2(5).failures;
cond_distr_params(15).beta = cond_distr_params(5).beta + observ_AV2(5).miles - observ_AV2(5).failures;
cond_distr_params(15).steps = cond_distr_params(5).steps;
alpha_obs_AV2 = alpha;
for i=1:length(alpha),
   alpha_obs_AV2(i) = alpha_obs_AV2(i) + observ_AV2(i).miles;
end
alpha_obs_AV2_fail = alpha_obs_AV2;

% AV3 - Update Beta parameters
cond_distr_params(16).partition = 1;
cond_distr_params(16).alfa = cond_distr_params(1).alfa + observ_AV3(1).failures;
cond_distr_params(16).beta = cond_distr_params(1).beta + observ_AV3(1).miles - observ_AV3(1).failures;
cond_distr_params(16).steps = cond_distr_params(1).steps;
cond_distr_params(17).partition = 2;
cond_distr_params(17).alfa = cond_distr_params(2).alfa + observ_AV3(2).failures;
cond_distr_params(17).beta = cond_distr_params(2).beta + observ_AV3(2).miles - observ_AV3(2).failures;
cond_distr_params(17).steps = cond_distr_params(2).steps;
cond_distr_params(18).partition = 3;
cond_distr_params(18).alfa = cond_distr_params(3).alfa + observ_AV3(3).failures;
cond_distr_params(18).beta = cond_distr_params(3).beta + observ_AV3(3).miles - observ_AV3(3).failures;
cond_distr_params(18).steps = cond_distr_params(3).steps;
cond_distr_params(19).partition = 4;
cond_distr_params(19).alfa = cond_distr_params(4).alfa + observ_AV3(4).failures;
cond_distr_params(19).beta = cond_distr_params(4).beta + observ_AV3(4).miles - observ_AV3(4).failures;
cond_distr_params(19).steps = cond_distr_params(4).steps;
cond_distr_params(20).partition = 5;
cond_distr_params(20).alfa = cond_distr_params(5).alfa + observ_AV3(5).failures;
cond_distr_params(20).beta = cond_distr_params(5).beta + observ_AV3(5).miles - observ_AV3(5).failures;
cond_distr_params(20).steps = cond_distr_params(5).steps;

alpha_obs_AV3 = alpha;
for i=1:length(alpha),
   alpha_obs_AV3(i) = alpha_obs_AV3(i) + observ_AV3(i).miles;
end

% AV4 - Update Beta parameters
cond_distr_params(21).partition = 1;
cond_distr_params(21).alfa = cond_distr_params(1).alfa + observ_AV4(1).failures;
cond_distr_params(21).beta = cond_distr_params(1).beta + observ_AV4(1).miles - observ_AV4(1).failures;
cond_distr_params(21).steps = cond_distr_params(1).steps;
cond_distr_params(22).partition = 2;
cond_distr_params(22).alfa = cond_distr_params(2).alfa + observ_AV4(2).failures;
cond_distr_params(22).beta = cond_distr_params(2).beta + observ_AV4(2).miles - observ_AV4(2).failures;
cond_distr_params(22).steps = cond_distr_params(2).steps;
cond_distr_params(23).partition = 3;
cond_distr_params(23).alfa = cond_distr_params(3).alfa + observ_AV4(3).failures;
cond_distr_params(23).beta = cond_distr_params(3).beta + observ_AV4(3).miles - observ_AV4(3).failures;
cond_distr_params(23).steps = cond_distr_params(3).steps;
cond_distr_params(24).partition = 4;
cond_distr_params(24).alfa = cond_distr_params(4).alfa + observ_AV4(4).failures;
cond_distr_params(24).beta = cond_distr_params(4).beta + observ_AV4(4).miles - observ_AV4(4).failures;
cond_distr_params(24).steps = cond_distr_params(4).steps;
cond_distr_params(25).partition = 5;
cond_distr_params(25).alfa = cond_distr_params(5).alfa + observ_AV4(5).failures;
cond_distr_params(25).beta = cond_distr_params(5).beta + observ_AV4(5).miles - observ_AV4(5).failures;
cond_distr_params(25).steps = cond_distr_params(5).steps;

alpha_obs_AV4 = alpha;
for i=1:length(alpha),
   alpha_obs_AV4(i) = alpha_obs_AV4(i) + observ_AV4(i).miles;
end
alpha_obs_AV4_fail = alpha_obs_AV4;

% AV5 - Update Beta parameters
cond_distr_params(26).partition = 1;
cond_distr_params(26).alfa = cond_distr_params(1).alfa + observ_AV5(1).failures;
cond_distr_params(26).beta = cond_distr_params(1).beta + observ_AV5(1).miles - observ_AV5(1).failures;
cond_distr_params(26).steps = cond_distr_params(1).steps;
cond_distr_params(27).partition = 2;
cond_distr_params(27).alfa = cond_distr_params(2).alfa + observ_AV5(2).failures;
cond_distr_params(27).beta = cond_distr_params(2).beta + observ_AV5(2).miles - observ_AV5(2).failures;
cond_distr_params(27).steps = cond_distr_params(2).steps;
cond_distr_params(28).partition = 3;
cond_distr_params(28).alfa = cond_distr_params(3).alfa + observ_AV5(3).failures;
cond_distr_params(28).beta = cond_distr_params(3).beta + observ_AV5(3).miles - observ_AV5(3).failures;
cond_distr_params(28).steps = cond_distr_params(3).steps;
cond_distr_params(29).partition = 4;
cond_distr_params(29).alfa = cond_distr_params(4).alfa + observ_AV5(4).failures;
cond_distr_params(29).beta = cond_distr_params(4).beta + observ_AV5(4).miles - observ_AV5(4).failures;
cond_distr_params(29).steps = cond_distr_params(4).steps;
cond_distr_params(30).partition = 5;
cond_distr_params(30).alfa = cond_distr_params(5).alfa + observ_AV5(5).failures;
cond_distr_params(30).beta = cond_distr_params(5).beta + observ_AV5(5).miles - observ_AV5(5).failures;
cond_distr_params(30).steps = cond_distr_params(5).steps;

alpha_obs_AV5 = alpha;
for i=1:length(alpha),
   alpha_obs_AV5(i) = alpha_obs_AV5(i) + observ_AV5(i).miles;
end
alpha_obs_AV5_fail = alpha_obs_AV5;

% AV Vendor  - Update Beta parameters
cond_distr_params(31).partition = 1;
cond_distr_params(31).alfa = cond_distr_params(1).alfa + observ_all(1).failures;
cond_distr_params(31).beta = cond_distr_params(1).beta + observ_all(1).miles - observ_all(1).failures;
cond_distr_params(31).steps = cond_distr_params(1).steps;
cond_distr_params(32).partition = 2;
cond_distr_params(32).alfa = cond_distr_params(2).alfa + observ_all(2).failures;
cond_distr_params(32).beta = cond_distr_params(2).beta + observ_all(2).miles - observ_all(2).failures;
cond_distr_params(32).steps = cond_distr_params(2).steps;
cond_distr_params(33).partition = 3;
cond_distr_params(33).alfa = cond_distr_params(3).alfa + observ_all(3).failures;
cond_distr_params(33).beta = cond_distr_params(3).beta + observ_all(3).miles - observ_all(3).failures;
cond_distr_params(33).steps = cond_distr_params(3).steps;
cond_distr_params(34).partition = 4;
cond_distr_params(34).alfa = cond_distr_params(4).alfa + observ_all(4).failures;
cond_distr_params(34).beta = cond_distr_params(4).beta + observ_all(4).miles - observ_all(4).failures;
cond_distr_params(34).steps = cond_distr_params(4).steps;
cond_distr_params(35).partition = 5;
cond_distr_params(35).alfa = cond_distr_params(5).alfa + observ_all(5).failures;
cond_distr_params(35).beta = cond_distr_params(5).beta + observ_all(5).miles - observ_all(5).failures;
cond_distr_params(35).steps = cond_distr_params(5).steps;

alpha_obs_all = alpha;
for i=1:length(alpha),
   alpha_obs_all(i) = alpha_obs_all(i) + observ_all(i).miles;
end

% AV3 - Update Beta() parameters for observations with failures
cond_distr_params(36).partition = 1;
cond_distr_params(36).alfa = cond_distr_params(1).alfa + observ_AV3_fail(1).failures;
cond_distr_params(36).beta = cond_distr_params(1).beta + observ_AV3_fail(1).miles - observ_AV3_fail(1).failures;
cond_distr_params(36).steps = cond_distr_params(1).steps;
cond_distr_params(37).partition = 2;
cond_distr_params(37).alfa = cond_distr_params(2).alfa + observ_AV3_fail(2).failures;
cond_distr_params(37).beta = cond_distr_params(2).beta + observ_AV3_fail(2).miles - observ_AV3_fail(2).failures;
cond_distr_params(37).steps = cond_distr_params(2).steps;
cond_distr_params(38).partition = 3;
cond_distr_params(38).alfa = cond_distr_params(3).alfa + observ_AV3_fail(3).failures;
cond_distr_params(38).beta = cond_distr_params(3).beta + observ_AV3_fail(3).miles - observ_AV3_fail(3).failures;
cond_distr_params(38).steps = cond_distr_params(3).steps;
cond_distr_params(39).partition = 4;
cond_distr_params(39).alfa = cond_distr_params(4).alfa + observ_AV3_fail(4).failures;
cond_distr_params(39).beta = cond_distr_params(4).beta + observ_AV3_fail(4).miles - observ_AV3_fail(4).failures;
cond_distr_params(39).steps = cond_distr_params(4).steps;
cond_distr_params(40).partition = 5;
cond_distr_params(40).alfa = cond_distr_params(5).alfa + observ_AV3_fail(5).failures;
cond_distr_params(40).beta = cond_distr_params(5).beta + observ_AV3_fail(5).miles - observ_AV3_fail(5).failures;
cond_distr_params(40).steps = cond_distr_params(5).steps;

alpha_obs_AV3_fail = alpha;
for i=1:length(alpha),
   alpha_obs_AV3_fail(i) = alpha_obs_AV3_fail(i) + observ_AV3_fail(i).miles;
end

% AV Vendor  - Update Beta parameters
cond_distr_params(41).partition = 1;
cond_distr_params(41).alfa = cond_distr_params(1).alfa + observ_all_fail(1).failures;
cond_distr_params(41).beta = cond_distr_params(1).beta + observ_all_fail(1).miles - observ_all_fail(1).failures;
cond_distr_params(41).steps = cond_distr_params(1).steps;
cond_distr_params(42).partition = 2;
cond_distr_params(42).alfa = cond_distr_params(2).alfa + observ_all_fail(2).failures;
cond_distr_params(42).beta = cond_distr_params(2).beta + observ_all_fail(2).miles - observ_all_fail(2).failures;
cond_distr_params(42).steps = cond_distr_params(2).steps;
cond_distr_params(43).partition = 3;
cond_distr_params(43).alfa = cond_distr_params(3).alfa + observ_all_fail(3).failures;
cond_distr_params(43).beta = cond_distr_params(3).beta + observ_all_fail(3).miles - observ_all_fail(3).failures;
cond_distr_params(43).steps = cond_distr_params(3).steps;
cond_distr_params(44).partition = 4;
cond_distr_params(44).alfa = cond_distr_params(4).alfa + observ_all_fail(4).failures;
cond_distr_params(44).beta = cond_distr_params(4).beta + observ_all_fail(4).miles - observ_all_fail(4).failures;
cond_distr_params(44).steps = cond_distr_params(4).steps;
cond_distr_params(45).partition = 5;
cond_distr_params(45).alfa = cond_distr_params(5).alfa + observ_all_fail(5).failures;
cond_distr_params(45).beta = cond_distr_params(5).beta + observ_all_fail(5).miles - observ_all_fail(5).failures;
cond_distr_params(45).steps = cond_distr_params(5).steps;

alpha_obs_all_fail = alpha;
for i=1:length(alpha),
   alpha_obs_all_fail(i) = alpha_obs_all_fail(i) + observ_all_fail(i).miles;
end

% Create placeholders for AV1 - AV5 posteriors with (fleet data from the
% vendor).


for i=6:45,     % generate the posterior Betas for the sets of observations (AV1 - AV5 + for the vendor). 
    % get the distrbutions of the conditional probabilities of failure
    cond_distr_params(i).pdf = gen_beta_m(cond_distr_params(i).alfa,cond_distr_params(i).beta,cond_distr_params(i).steps);
end

% Now calculate the posterior Dirichlet distributions for the observations
% (AV1 - AV5 + Vendor)
integral_obs_AV1 = double(0);
integral_obs_AV1_fail = double(0);
integral_obs_AV2 = double(0);
integral_obs_AV3 = double(0);
integral_obs_AV3_fail = double(0);
integral_obs_AV4 = double(0);
integral_obs_AV5 = double(0);
integral_obs_all = double(0);
integral_obs_all_fail = double(0);
ind_space_obs_AV1 = ind_space;
ind_space_obs_AV1_fail = ind_space;
ind_space_obs_AV2 = ind_space;
ind_space_obs_AV3 = ind_space;
ind_space_obs_AV3_fail = ind_space;     % work with AV3 with failures
ind_space_obs_AV4 = ind_space;
ind_space_obs_AV5 = ind_space;
ind_space_obs_all = ind_space;      
ind_space_obs_all_fail = ind_space;     % work with Vendor with failures
for i=1:count,
    inds_1 = ind_space_obs_AV1(i).indeces;
    inds_1_fail = ind_space_obs_AV1_fail(i).indeces;
    inds_2 = ind_space_obs_AV2(i).indeces;
    inds_3 = ind_space_obs_AV3(i).indeces;
    inds_3_fail = ind_space_obs_AV3_fail(i).indeces;
    inds_4 = ind_space_obs_AV4(i).indeces;
    inds_5 = ind_space_obs_AV5(i).indeces;
    inds_all = ind_space_obs_all(i).indeces;
    inds_all_fail = ind_space_obs_all_fail(i).indeces;
    x_1 = double(inds_1) / double(steps);
    x_1_fail = double(inds_1_fail) / double(steps);
    x_2 = double(inds_2) / double(steps);
    x_3 = double(inds_3) / double(steps);
    x_3_fail = double(inds_3_fail) / double(steps);
    x_4 = double(inds_4) / double(steps);
    x_5 = double(inds_5) / double(steps);
    x_all = double(inds_all) / double(steps);
    x_all_fail = double(inds_all_fail) / double(steps);
    display(x_1);
    display(x_1_fail);
    display(x_2);
    display(x_3);
    display(x_3_fail);
    display(x_4);
    display(x_5);
    display(x_all);
    display(x_all_fail);
    ind_space_obs_AV1(i).x = x_1;
    ind_space_obs_AV1_fail(i).x = x_1_fail;
    ind_space_obs_AV2(i).x = x_2;
    ind_space_obs_AV3(i).x = x_3;
    ind_space_obs_AV3_fail(i).x = x_3_fail;
    ind_space_obs_AV4(i).x = x_4;
    ind_space_obs_AV5(i).x = x_5;
    ind_space_obs_all(i).x = x_all;
    ind_space_obs_all_fail(i).x = x_all_fail;
    D_obs1 = dirichlet_robust(x_1,alpha_obs_AV1);
    D_obs1_fail = dirichlet_robust(x_1_fail,alpha_obs_AV1_fail);
    D_obs2 = dirichlet_robust(x_2,alpha_obs_AV2);
    D_obs3 = dirichlet_robust(x_3,alpha_obs_AV3);
    D_obs3_fail = dirichlet_robust(x_3_fail,alpha_obs_AV3_fail);
    D_obs4 = dirichlet_robust(x_4,alpha_obs_AV4);
    D_obs5 = dirichlet_robust(x_5,alpha_obs_AV5);
    D_obs_all = dirichlet_robust(x_all,alpha_obs_all);
    D_obs_all_fail = dirichlet_robust(x_all_fail,alpha_obs_all_fail);
    ind_space_obs_AV1(i).pdf = D_obs1;
    ind_space_obs_AV1_fail(i).pdf = D_obs1_fail;
    ind_space_obs_AV2(i).pdf = D_obs2;
    ind_space_obs_AV3(i).pdf = D_obs3;
    ind_space_obs_AV3_fail(i).pdf = D_obs3_fail;
    ind_space_obs_AV4(i).pdf = D_obs4;
    ind_space_obs_AV5(i).pdf = D_obs5;
    ind_space_obs_all(i).pdf = D_obs_all;
    ind_space_obs_all_fail(i).pdf = D_obs_all_fail;
    display(D_obs1);
    display(D_obs1_fail);
    display(D_obs2);
    display(D_obs3);
    display(D_obs3_fail);
    display(D_obs4);
    display(D_obs5);
    display(D_obs_all);
    display(D_obs_all_fail);
    ind_space_obs_AV1(i).volume = D_obs1 * vol_unit;
    integral_obs_AV1 = integral_obs_AV1 + D_obs1 * vol_unit;
    ind_space_obs_AV1_fail(i).volume = D_obs1_fail * vol_unit;
    integral_obs_AV1_fail = integral_obs_AV1_fail + D_obs1_fail * vol_unit;
    ind_space_obs_AV2(i).volume = D_obs2 * vol_unit;
    integral_obs_AV2 = integral_obs_AV2 + D_obs2 * vol_unit;
    ind_space_obs_AV3(i).volume = D_obs3 * vol_unit;
    integral_obs_AV3 = integral_obs_AV3 + D_obs3 * vol_unit;
    ind_space_obs_AV3_fail(i).volume = D_obs3_fail * vol_unit;
    integral_obs_AV3_fail = integral_obs_AV3_fail + D_obs3_fail * vol_unit;
    ind_space_obs_AV4(i).volume = D_obs4 * vol_unit;
    integral_obs_AV4 = integral_obs_AV4 + D_obs4 * vol_unit;
    ind_space_obs_AV5(i).volume = D_obs5 * vol_unit;
    integral_obs_AV5 = integral_obs_AV5 + D_obs5 * vol_unit;
    ind_space_obs_all(i).volume = D_obs_all * vol_unit;
    integral_obs_all = integral_obs_all + D_obs_all * vol_unit;
    ind_space_obs_all_fail(i).volume = D_obs_all_fail * vol_unit;
    integral_obs_all_fail = integral_obs_all_fail + D_obs_all_fail * vol_unit;
end
display(integral_obs_AV1);
display(integral_obs_AV1_fail);
display(integral_obs_AV2);
display(integral_obs_AV3);
display(integral_obs_AV3_fail);
display(integral_obs_AV4);
display(integral_obs_AV5);
display(integral_obs_all);
display(integral_obs_all_fail);
integral_2_obs_AV1 = double(0);
integral_2_obs_AV1_fail = double(0);
integral_2_obs_AV2 = double(0);
integral_2_obs_AV3 = double(0);
integral_2_obs_AV3_fail = double(0);
integral_2_obs_AV4 = double(0);
integral_2_obs_AV5 = double(0);
integral_2_obs_all = double(0);
integral_2_obs_all_fail = double(0);
for i=1:count,
 ind_space_obs_AV1(i).volume = ind_space_obs_AV1(i).volume / integral_obs_AV1;   % compensate rounding errors and set the integral to 1. 
 integral_2_obs_AV1 = integral_2_obs_AV1 + ind_space_obs_AV1(i).volume;
 ind_space_obs_AV1_fail(i).volume = ind_space_obs_AV1_fail(i).volume / integral_obs_AV1_fail;   % compensate rounding errors and set the integral to 1. 
 integral_2_obs_AV1_fail = integral_2_obs_AV1_fail + ind_space_obs_AV1_fail(i).volume;
 ind_space_obs_AV2(i).volume = ind_space_obs_AV2(i).volume / integral_obs_AV2;   % compensate rounding errors and set the integral to 1. 
 integral_2_obs_AV2 = integral_2_obs_AV2 + ind_space_obs_AV2(i).volume;
 ind_space_obs_AV3(i).volume = ind_space_obs_AV3(i).volume / integral_obs_AV3;   % compensate rounding errors and set the integral to 1. 
 integral_2_obs_AV3 = integral_2_obs_AV3 + ind_space_obs_AV3(i).volume;
 ind_space_obs_AV3_fail(i).volume = ind_space_obs_AV3_fail(i).volume / integral_obs_AV3_fail;   % compensate rounding errors and set the integral to 1. 
 integral_2_obs_AV3_fail = integral_2_obs_AV3_fail + ind_space_obs_AV3_fail(i).volume;
 ind_space_obs_AV4(i).volume = ind_space_obs_AV4(i).volume / integral_obs_AV4;   % compensate rounding errors and set the integral to 1. 
 integral_2_obs_AV4 = integral_2_obs_AV4 + ind_space_obs_AV4(i).volume;
 ind_space_obs_AV5(i).volume = ind_space_obs_AV4(i).volume / integral_obs_AV5;   % compensate rounding errors and set the integral to 1. 
 integral_2_obs_AV5 = integral_2_obs_AV5 + ind_space_obs_AV5(i).volume;
 ind_space_obs_all(i).volume = ind_space_obs_all(i).volume / integral_obs_all;   % compensate rounding errors and set the integral to 1. 
 integral_2_obs_all = integral_2_obs_all + ind_space_obs_all(i).volume;
 ind_space_obs_all_fail(i).volume = ind_space_obs_all_fail(i).volume / integral_obs_all_fail;   % compensate rounding errors and set the integral to 1. 
 integral_2_obs_all_fail = integral_2_obs_all_fail + ind_space_obs_all_fail(i).volume;
end
display('integral_2_obs_AV1 = ')
display(integral_2_obs_AV1);
display('integral_2_obs_AV1_fail = ')
display(integral_2_obs_AV1_fail);
display('integral_2_obs_AV2 = ')
display(integral_2_obs_AV2);
display('integral_2_obs_AV3 = ')
display(integral_2_obs_AV3);
display('integral_2_obs_AV3_fail = ')
display(integral_2_obs_AV3_fail);
display('integral_2_obs_AV4 = ')
display(integral_2_obs_AV4);
display('integral_2_obs_AV5 = ')
display(integral_2_obs_AV5);
display('integral_2_obs_all = ')
display(integral_2_obs_all);
display('integral_2_obs_all_fail = ')
display(integral_2_obs_all_fail);

% Now for AV1 - AV5 need to compute the posteriors for the miles/failures 
% observed by the vendor. 
% Best to use the conditional distributions computed for the vendor when 
% computing the marginal posterior (fleet data).  

% Now get the marginal posterior distributions
post_marg_1 = Lin_Mat(cond_distr_params(1).steps);          % White box inference AV1 (own data only)
post_marg_1_full = Lin_Mat(cond_distr_params(1).steps);     % White box inference AV1 (fleet data)
post_marg_1_full_fail = Lin_Mat(cond_distr_params(1).steps);     % White box inference AV1 (fleet data)
post_marg_2 = Lin_Mat(cond_distr_params(1).steps);          % White box inference AV2 (own data only)
post_marg_2_full = Lin_Mat(cond_distr_params(1).steps);     % White box inference AV2 (fleet data)
post_marg_3 = Lin_Mat(cond_distr_params(1).steps);          % White box inference AV3 (own data only)
post_marg_3_full = Lin_Mat(cond_distr_params(1).steps);     % White box inference AV3 (fleet data)
post_marg_3_fail = Lin_Mat(cond_distr_params(1).steps);     % White box inference AV3 fail (own data only)
post_marg_3_full_fail = Lin_Mat(cond_distr_params(1).steps);% White box inference AV3 fail (fleet data)
post_marg_4 = Lin_Mat(cond_distr_params(1).steps);          % White box inference AV4 (own data only)
post_marg_4_full = Lin_Mat(cond_distr_params(1).steps);     % White box inference AV4 (fleet data)
post_marg_5 = Lin_Mat(cond_distr_params(1).steps);          % White box inference AV5 (own data only)
post_marg_5_full = Lin_Mat(cond_distr_params(1).steps);     % White box inference AV5 (fleet data)
post_marg_all = Lin_Mat(cond_distr_params(1).steps);        % White box inference Vendor (fleet data )
post_marg_all_fail = Lin_Mat(cond_distr_params(1).steps);   % White box inference Vendor fail (fleet data )

% Need to update the marginal distributions. 

for i=1:count,
    for j=1:dims,
        c_dist_1 = gen_beta_scaled(cond_distr_params(j+5).alfa,cond_distr_params(j+5).beta,ind_space_obs_AV1(i).x(j),cond_distr_params(j+5).steps);
        c_dist_2 = gen_beta_scaled(cond_distr_params(j+10).alfa,cond_distr_params(j+10).beta,ind_space_obs_AV2(i).x(j),cond_distr_params(j+10).steps);
        c_dist_3 = gen_beta_scaled(cond_distr_params(j+15).alfa,cond_distr_params(j+15).beta,ind_space_obs_AV3(i).x(j),cond_distr_params(j+15).steps);
        c_dist_3_fail = gen_beta_scaled(cond_distr_params(j+15).alfa,cond_distr_params(j+15).beta,ind_space_obs_AV3_fail(i).x(j),cond_distr_params(j+15).steps);
        c_dist_4 = gen_beta_scaled(cond_distr_params(j+20).alfa,cond_distr_params(j+20).beta,ind_space_obs_AV4(i).x(j),cond_distr_params(j+20).steps);
        c_dist_5 = gen_beta_scaled(cond_distr_params(j+25).alfa,cond_distr_params(j+25).beta,ind_space_obs_AV5(i).x(j),cond_distr_params(j+25).steps);
        c_dist_all = gen_beta_scaled(cond_distr_params(j+30).alfa,cond_distr_params(j+30).beta,ind_space_obs_all(i).x(j),cond_distr_params(j+30).steps);
        c_dist_all_fail = gen_beta_scaled(cond_distr_params(j+30).alfa,cond_distr_params(j+30).beta,ind_space_obs_all_fail(i).x(j),cond_distr_params(j+30).steps);
        c_dist_1_full = c_dist_all;     % AV1 with fleet data
        c_dist_1_full_fail = c_dist_all_fail;     % AV1 with fleet data
        c_dist_2_full = c_dist_all;     % AV2 with fleet data
        c_dist_3_full = c_dist_all;     % AV3 with fleet data
        c_dist_3_full_fail = c_dist_all_fail;     % AV3 with fleet data
        c_dist_4_full = c_dist_all;     % AV4 with fleet data
        c_dist_5_full = c_dist_all;     % AV5 with fleet data        
        if(j==1)
            post_mg_1 = c_dist_1;     % marg_conv 
            post_mg_2 = c_dist_2;
            post_mg_3 = c_dist_3;
            post_mg_3_fail = c_dist_3_fail;
            post_mg_4 = c_dist_4;
            post_mg_5 = c_dist_5;
            post_mg_all = c_dist_all;
            post_mg_all_fail = c_dist_all_fail;
            post_mg_1_full = c_dist_1_full;
            post_mg_1_full_fail = c_dist_1_full_fail;
            post_mg_2_full = c_dist_2_full;
            post_mg_3_full = c_dist_3_full;
            post_mg_3_full_fail = c_dist_3_full_fail;
            post_mg_4_full = c_dist_4_full;
            post_mg_5_full = c_dist_5_full;
        else
            post_mg_1 = conv(post_mg_1,c_dist_1);
            post_mg_1(length(post_mg_1)+1) = double(0);
            post_mg_2 = conv(post_mg_2,c_dist_2);
            post_mg_2(length(post_mg_2)+1) = double(0);
            post_mg_3 = conv(post_mg_3,c_dist_3);
            post_mg_3(length(post_mg_3)+1) = double(0);
            post_mg_3_fail = conv(post_mg_3_fail,c_dist_3_fail);
            post_mg_3_fail(length(post_mg_3_fail)+1) = double(0);
            post_mg_4 = conv(post_mg_4,c_dist_4);
            post_mg_4(length(post_mg_4)+1) = double(0);
            post_mg_5 = conv(post_mg_5,c_dist_5);
            post_mg_5(length(post_mg_5)+1) = double(0);
            post_mg_all = conv(post_mg_all,c_dist_all);
            post_mg_all(length(post_mg_all)+1) = double(0);
            post_mg_all_fail = conv(post_mg_all_fail,c_dist_all_fail);
            post_mg_all_fail(length(post_mg_all_fail)+1) = double(0);
            post_mg_1_full = conv(post_mg_1_full,c_dist_1_full);
            post_mg_1_full(length(post_mg_1_full)+1) = double(0);
            post_mg_1_full_fail = conv(post_mg_1_full_fail,c_dist_1_full_fail);
            post_mg_1_full_fail(length(post_mg_1_full_fail)+1) = double(0);
            post_mg_2_full = conv(post_mg_2_full,c_dist_2_full);
            post_mg_2_full(length(post_mg_2_full)+1) = double(0);
            post_mg_3_full = conv(post_mg_3_full,c_dist_3_full);
            post_mg_3_full(length(post_mg_3_full)+1) = double(0);
            post_mg_3_full_fail = conv(post_mg_3_full_fail,c_dist_3_full_fail);
            post_mg_3_full_fail(length(post_mg_3_full_fail)+1) = double(0);
            post_mg_4_full = conv(post_mg_4_full,c_dist_4_full);
            post_mg_4_full(length(post_mg_4_full)+1) = double(0);
            post_mg_5_full = conv(post_mg_5_full,c_dist_5_full);
            post_mg_5_full(length(post_mg_5_full)+1) = double(0);
        end
    end
    % now apply the weigth to this convolution
    post_mg_1 = post_mg_1 * ind_space_obs_AV1(i).volume; 
    post_mg_2 = post_mg_2 * ind_space_obs_AV2(i).volume; 
    post_mg_3 = post_mg_3 * ind_space_obs_AV3(i).volume; 
    post_mg_3_fail = post_mg_3_fail * ind_space_obs_AV3_fail(i).volume; 
    post_mg_4 = post_mg_4 * ind_space_obs_AV4(i).volume; 
    post_mg_5 = post_mg_5 * ind_space_obs_AV5(i).volume; 
    post_mg_all = post_mg_all * ind_space_obs_all(i).volume; 
    post_mg_all_fail = post_mg_all_fail * ind_space_obs_all_fail(i).volume; 
    post_mg_1_full = post_mg_1_full * ind_space_obs_all(i).volume; 
    post_mg_1_full_fail = post_mg_1_full_fail * ind_space_obs_all_fail(i).volume; 
    post_mg_2_full = post_mg_2_full * ind_space_obs_all(i).volume; 
    post_mg_3_full = post_mg_3_full * ind_space_obs_all(i).volume; 
    post_mg_3_full_fail = post_mg_3_full_fail * ind_space_obs_all_fail(i).volume; 
    post_mg_4_full = post_mg_4_full * ind_space_obs_all(i).volume; 
    post_mg_5_full = post_mg_5_full * ind_space_obs_all(i).volume; 
    % add add it to the marginal distribution
    post_marg_1 = post_marg_1 + post_mg_1;
    post_marg_2 = post_marg_2 + post_mg_2;
    post_marg_3 = post_marg_3 + post_mg_3;
    post_marg_3_fail = post_marg_3_fail + post_mg_3_fail;
    post_marg_4 = post_marg_4 + post_mg_4;
    post_marg_5 = post_marg_5 + post_mg_5;
    post_marg_all = post_marg_all + post_mg_all;
    post_marg_all_fail = post_marg_all_fail + post_mg_all_fail;
    post_marg_1_full = post_marg_1_full + post_mg_1_full;
    post_marg_1_full_fail = post_marg_1_full_fail + post_mg_1_full_fail;
    post_marg_2_full = post_marg_2_full + post_mg_2_full;
    post_marg_3_full = post_marg_3_full + post_mg_3_full;
    post_marg_3_full_fail = post_marg_3_full_fail + post_mg_3_full_fail;
    post_marg_4_full = post_marg_4_full + post_mg_4_full;
    post_marg_5_full = post_marg_5_full + post_mg_5_full;
end

display(sum(post_marg_1));
display(sum(post_marg_2));
display(sum(post_marg_3));
display(sum(post_marg_3_fail));
display(sum(post_marg_4));
display(sum(post_marg_5));
display(sum(post_marg_all));
display(sum(post_marg_all_fail));
display(sum(post_marg_1_full));
display(sum(post_marg_1_full_fail));
display(sum(post_marg_2_full));
display(sum(post_marg_3_full));
display(sum(post_marg_3_full_fail));
display(sum(post_marg_4_full));
display(sum(post_marg_5_full));

% May need to normalise the marginal posteriors.

% Now need to conduct black-box inferences for AV1 - AV5 and for the
% vendor. AV 1 - AV5 will compute these using own observations only. 
% AV1 structures
plot_array_descr(20).Plot.Y = marginal_distr;      % Should be set to the marginal prior.
plot_array_descr(21).Plot.Y = marginal_distr;      % Black-box posterior (oun data)
plot_array_descr(21) = BlackBoxBayesianInference(plot_array_descr(21), total_failures(1), total_miles(1));
% Modify LD of the black-box 
% plot_array_descr(21).Plot.LD = 'BB marginal posterior PFM (AV1: own data)';
plot_array_descr(22).Plot.Y = marginal_distr;      % Black-box posterior (fleet data)
plot_array_descr(22) = BlackBoxBayesianInference(plot_array_descr(22), vendor_total_failures, vendor_total_miles);
% Modify LD of the black-box 
% plot_array_descr(22).Plot.LD = 'BB marginal posterior PFM (AV1: fleet data)';
plot_array_descr(23).Plot.Y = post_marg_1;         % White-box posterior (oun data)
plot_array_descr(24).Plot.Y = post_marg_1_full;    % White-box posterior (fleet data)

% AV2 structures
plot_array_descr(25).Plot.Y = marginal_distr;      % Should be set to the marginal prior.
plot_array_descr(26).Plot.Y = marginal_distr;      % Black-box posterior (oun data)
plot_array_descr(26) = BlackBoxBayesianInference(plot_array_descr(26), total_failures(2), total_miles(2));
% Modify LD of the black-box 
% plot_array_descr(26).Plot.LD = 'BB marginal posterior PFM (AV2: own data)';
plot_array_descr(27).Plot.Y = marginal_distr;      % Black-box posterior (fleet data)
plot_array_descr(27) = BlackBoxBayesianInference(plot_array_descr(27), vendor_total_failures, vendor_total_miles);
% Modify LD of the black-box 
% plot_array_descr(27).Plot.LD = 'BB marginal posterior PFM (AV2: fleet data)';
plot_array_descr(28).Plot.Y = post_marg_2;         % White-box posterior (oun data)
plot_array_descr(29).Plot.Y = post_marg_2_full;    % White-box posterior (fleet data)

% AV3 structures
plot_array_descr(30).Plot.Y = marginal_distr;      % Should be set to the marginal prior.
plot_array_descr(31).Plot.Y = marginal_distr;      % Black-box posterior (oun data)
plot_array_descr(31) = BlackBoxBayesianInference(plot_array_descr(31), total_failures(3), total_miles(3));
% Modify LD of the black-box 
% plot_array_descr(31).Plot.LD = 'BB marginal posterior PFM (AV3: own data)';
plot_array_descr(32).Plot.Y = marginal_distr;      % Black-box posterior (fleet data)
plot_array_descr(32) = BlackBoxBayesianInference(plot_array_descr(32), vendor_total_failures, vendor_total_miles);
% Modify LD of the black-box 
% plot_array_descr(32).Plot.LD = 'BB marginal posterior PFM (AV3: fleet data)';
plot_array_descr(33).Plot.Y = post_marg_3;         % White-box posterior (oun data)
plot_array_descr(34).Plot.Y = post_marg_3_full;    % White-box posterior (fleet data)

% AV4 structures
plot_array_descr(35).Plot.Y = marginal_distr;      % Should be set to the marginal prior.
plot_array_descr(36).Plot.Y = marginal_distr;      % Black-box posterior (oun data)
plot_array_descr(36) = BlackBoxBayesianInference(plot_array_descr(36), total_failures(4), total_miles(4));
% Modify LD of the black-box 
% plot_array_descr(36).Plot.LD = 'BB marginal posterior PFM (AV4: own data)';
plot_array_descr(37).Plot.Y = marginal_distr;      % Black-box posterior (fleet data)
plot_array_descr(37) = BlackBoxBayesianInference(plot_array_descr(37), vendor_total_failures, vendor_total_miles);
% Modify LD of the black-box 
% plot_array_descr(37).Plot.LD = 'BB marginal posterior PFM (AV4: fleet data)';
plot_array_descr(38).Plot.Y = post_marg_4;         % White-box posterior (oun data)
plot_array_descr(39).Plot.Y = post_marg_4_full;    % White-box posterior (fleet data)

% AV5 structures
plot_array_descr(40).Plot.Y = marginal_distr;      % Should be set to the marginal prior.
plot_array_descr(41).Plot.Y = marginal_distr;      % Black-box posterior (oun data)
plot_array_descr(41) = BlackBoxBayesianInference(plot_array_descr(41), total_failures(5), total_miles(5));
% Modify LD of the black-box 
% plot_array_descr(41).Plot.LD = 'BB marginal posterior PFM (AV5: own data)';
plot_array_descr(42).Plot.Y = marginal_distr;      % Black-box posterior (fleet data)
plot_array_descr(42) = BlackBoxBayesianInference(plot_array_descr(42), vendor_total_failures, vendor_total_miles);
% Modify LD of the black-box 
% plot_array_descr(43).Plot.Plot.LD = 'BB marginal posterior PFM (AV5: fleet data)';
plot_array_descr(43).Plot.Y = post_marg_5;         % White-box posterior (own data)
plot_array_descr(44).Plot.Y = post_marg_5_full;    % White-box posterior (fleet data)

% Vendor structures
plot_array_descr(45).Plot.Y = marginal_distr;      % Should be set to the marginal prior.
plot_array_descr(46).Plot.Y = marginal_distr;      % Black-box posterior (fleet data)
plot_array_descr(46) = BlackBoxBayesianInference(plot_array_descr(46), vendor_total_failures, vendor_total_miles);
% Modify LD of the black-box 
% plot_array_descr(47).Plot.LD = 'WB marginal posterior PFM (Vendor: fleet data)';
plot_array_descr(47).Plot.Y = post_marg_all;       % white-box posterior (fleet data)

% Results with failures
% AV1 48 - BB, 49 - WB
plot_array_descr(48).Plot.Y = marginal_distr;      % Should be set to the marginal prior.
plot_array_descr(48) = BlackBoxBayesianInference(plot_array_descr(48), vendor_total_failures_fail, vendor_total_miles_fail);
% Modify LD of the black-box 
% plot_array_descr(48).Plot.LD = 'BB marginal posterior PFM (AV1: fleet data with failures)';
plot_array_descr(49).Plot.Y = post_marg_1_full_fail;    % White-box posterior (fleet data)

% AV3 52 - BB (own), 53 - BB (all), 54 - WB (own), 55 - WB (all)
plot_array_descr(52).Plot.Y = marginal_distr;      % Should be set to the marginal prior.
plot_array_descr(52) = BlackBoxBayesianInference(plot_array_descr(31), total_failures(3), total_miles(3));
% Modify LD of the black-box 
% plot_array_descr(52).Plot.LD = 'BB marginal posterior PFM (AV2: own data with failures)';
plot_array_descr(53).Plot.Y = marginal_distr;      % Black-box posterior (fleet data)
plot_array_descr(53) = BlackBoxBayesianInference(plot_array_descr(53), vendor_total_failures_fail, vendor_total_miles_fail);
% Modify LD of the black-box 
% plot_array_descr(53).Plot.LD = 'BB marginal posterior PFM (AV3: fleet data with failures)';
% Modify LD of the black-box 
plot_array_descr(54).Plot.Y = post_marg_3_fail;         % White-box posterior (own data)
% plot_array_descr(54).Plot.LD = 'WB marginal posterior PFM (AV3: own data with failures)';
% Modify LD of the black-box 
plot_array_descr(55).Plot.Y = post_marg_3_full_fail;    % White-box posterior (fleet data)
% plot_array_descr(55).Plot.LD = 'WB marginal posterior PFM (AV3: fleet data with failures)';

% Vendor 60 - BB, 61 - WB
plot_array_descr(60).Plot.Y = marginal_distr;      % Black-box posterior (fleet data)
plot_array_descr(60) = BlackBoxBayesianInference(plot_array_descr(60), vendor_total_failures_fail, vendor_total_miles_fail);
% Modify LD of the black-box 
% plot_array_descr(60).Plot.LD = 'BB marginal posterior PFM (Vendor: fleet data with failures)';
plot_array_descr(61).Plot.Y = post_marg_all_fail;       % white-box posterior (fleet data)

clear Data;

% Pring the results
clear Data;
P60(1) = plot_array_descr(20);      % Marginal Prior AV1
P60(2) = plot_array_descr(21);      % BB marginal posterior AV1 (own data)
P60(3) = plot_array_descr(22);      % BB marginal posterior AV1 (fleet data)
P60(4) = plot_array_descr(23);      % WB marginal posterior AV1 (own data)
P60(5) = plot_array_descr(24);      % WB marginal posterior AV1 (fleet data)
Data(1).Plot = P60;
Data(1).xlabel = P60(1).Plot.XLabel;
Data(1).ylabel = P60(1).Plot.YLabel;
Data(1).xlim = 0.01;
Data(1).title = 'AV1: White-box vs Black-box inference';

P61(1) = plot_array_descr(25);      % Marginal Prior AV2
P61(2) = plot_array_descr(26);      % BB marginal posterior AV2 (own data)
P61(3) = plot_array_descr(27);      % BB marginal posterior AV2 (fleet data)
P61(4) = plot_array_descr(28);      % WB marginal posterior AV2 (own data)
P61(5) = plot_array_descr(29);      % WB marginal posterior AV2 (fleet data)
Data(2).Plot = P61;
Data(2).xlabel = P61(1).Plot.XLabel;
Data(2).ylabel = P61(1).Plot.YLabel;
Data(2).xlim = 0.01;
Data(2).title = 'AV2: White-box vs Black-box inference';

P62(1) = plot_array_descr(30);      % Marginal Prior AV3
P62(2) = plot_array_descr(31);      % BB marginal posterior AV3 (own data)
P62(3) = plot_array_descr(32);      % BB marginal posterior AV3 (fleet data)
P62(4) = plot_array_descr(33);      % WB marginal posterior AV3 (own data)
P62(5) = plot_array_descr(34);      % WB marginal posterior AV3 (fleet data)
Data(3).Plot = P62;
Data(3).xlabel = P62(1).Plot.XLabel;
Data(3).ylabel = P62(1).Plot.YLabel;
Data(3).xlim = 0.01;
Data(3).title = 'AV3: White-box vs Black-box inference';

P63(1) = plot_array_descr(35);      % Marginal Prior AV4
P63(2) = plot_array_descr(36);      % BB marginal posterior AV4 (own data)
P63(3) = plot_array_descr(37);      % BB marginal posterior AV4 (fleet data)
P63(4) = plot_array_descr(38);      % WB marginal posterior AV4 (own data)
P63(5) = plot_array_descr(39);      % WB marginal posterior AV4 (fleet data)
Data(4).Plot = P63;
Data(4).xlabel = P63(1).Plot.XLabel;
Data(4).ylabel = P63(1).Plot.YLabel;
Data(4).xlim = 0.01;
Data(4).title = 'AV4: White-box vs Black-box inference';

P64(1) = plot_array_descr(40);      % Marginal Prior AV5
P64(2) = plot_array_descr(41);      % BB marginal posterior AV4 (own data)
P64(3) = plot_array_descr(42);      % BB marginal posterior AV4 (fleet data)
P64(4) = plot_array_descr(43);      % WB marginal posterior AV4 (own data)
P64(5) = plot_array_descr(44);      % WB marginal posterior AV4 (fleet data)
Data(5).Plot = P64;
Data(5).xlabel = P64(1).Plot.XLabel;
Data(5).ylabel = P64(1).Plot.YLabel;
Data(5).xlim = 0.015;
Data(5).title = 'AV5: White-box vs Black-box inference';

P65(1) = plot_array_descr(45);      % Marginal Prior Vendor
P65(2) = plot_array_descr(46);      % BB marginal posterior Vendor
P65(3) = plot_array_descr(47);      % WB marginal posterior Vendor
Data(6).Plot = P65;
Data(6).xlabel = P65(1).Plot.XLabel;
Data(6).ylabel = P65(1).Plot.YLabel;
Data(6).xlim = 0.01;
Data(6).title = 'Vendor: White-box vs Black-box inference';

[fig_num,subplot_num] = plotting_subplots(30,Data,1,3,2);

clear Data;
P70(1) = plot_array_descr(20);      % Marginal Prior AV1
P70(2) = plot_array_descr(21);      % BB marginal posterior AV1 (own data)
P70(3) = plot_array_descr(22);      % BB marginal posterior AV1 (fleet data)
P70(4) = plot_array_descr(23);      % WB marginal posterior AV1 (own data)
P70(5) = plot_array_descr(24);      % WB marginal posterior AV1 (fleet data)
P70(6) = plot_array_descr(48);      % BB marginal posterior AV1 (fleet data, failures)
P70(7) = plot_array_descr(49);      % WB marginal posterior AV1 (fleet data, failures)

Data(1).Plot = P70;
Data(1).xlabel = P70(1).Plot.XLabel;
Data(1).ylabel = P70(1).Plot.YLabel;
Data(1).xlim = 0.01;
Data(1).title = 'AV1: White-box vs Black-box, fleet failures';

P71(1) = plot_array_descr(30);      % Marginal Prior AV3
P71(2) = plot_array_descr(31);      % BB marginal posterior AV3 (own data)
P71(3) = plot_array_descr(32);      % BB marginal posterior AV3 (fleet data)
P71(4) = plot_array_descr(33);      % WB marginal posterior AV3 (own data)
P71(5) = plot_array_descr(34);      % WB marginal posterior AV3 (fleet data)
P71(6) = plot_array_descr(52);      % BB marginal posterior AV3 (own data, failures)
P71(7) = plot_array_descr(53);      % BB marginal posterior AV3 (fleet data, failures)
P71(8) = plot_array_descr(54);      % WB marginal posterior AV3 (own data, failures)
P71(9) = plot_array_descr(55);      % WB marginal posterior AV3 (fleet data, failures)

Data(2).Plot = P71;
Data(2).xlabel = P71(1).Plot.XLabel;
Data(2).ylabel = P71(1).Plot.YLabel;
Data(2).xlim = 0.01;
Data(2).title = 'AV3: White-box vs Black-box, own/fleet failures';

P72(1) = plot_array_descr(45);      % Marginal Prior Vendor
P72(2) = plot_array_descr(46);      % BB marginal posterior Vendor
P72(3) = plot_array_descr(47);      % BB marginal posterior Vendor
P72(4) = plot_array_descr(60);      % BB marginal posterior Vendor (failures)
P72(5) = plot_array_descr(61);      % WB marginal posterior Vendor (failures)

Data(3).Plot = P72;
Data(3).xlabel = P72(1).Plot.XLabel;
Data(3).ylabel = P72(1).Plot.YLabel;
Data(3).xlim = 0.01;
Data(3).title = 'Vendor: White-box vs Black-box fleet failures';

P73(1) = plot_array_descr(20);      % Marginal Prior (bootstraping)
P73(2) = plot_array_descr(21);      % AV1 BB marginal posterior (own data, failures). Fleet failures do not affect posterior. 
P73(3) = plot_array_descr(48);      % AV1 BB marginal posterior (fleet data, failures)
P73(4) = plot_array_descr(23);      % AV1 WB marginal posterior (own data, fleet failures). Fleet failures do not affect posterior.
P73(5) = plot_array_descr(49);      % AV1 WB marginal posterior (fleet data, failures)

Data(4).Plot = P73;
Data(4).xlabel = P73(1).Plot.XLabel;
Data(4).ylabel = P73(1).Plot.YLabel;
Data(4).xlim = 0.01;
Data(4).title = 'AV1: White-box vs Black-box with fleet failures';

P74(1) = plot_array_descr(20);      % Marginal Prior (bootstraping)
P74(2) = plot_array_descr(52);      % AV3 BB marginal posterior (own data, failures). 
P74(3) = plot_array_descr(53);      % AV3 BB marginal posterior (fleet data, failures)
P74(4) = plot_array_descr(54);      % AV3 WB marginal posterior (own data, fleet failures). Fleet failures do not affect posterior.
P74(5) = plot_array_descr(55);      % AV3 WB marginal posterior (fleet data, failures)

Data(5).Plot = P74;
Data(5).xlabel = P74(1).Plot.XLabel;
Data(5).ylabel = P74(1).Plot.YLabel;
Data(5).xlim = 0.01;
Data(5).title = 'AV3: White-box vs Black-box with own failures';

P75(1) = plot_array_descr(20);      % Marginal Prior (bootstraping)
P75(2) = plot_array_descr(60);      % Vendor BB marginal posterior (fleet failures). 
P75(3) = plot_array_descr(61);      % Vendor WB marginal posterior (fleet failures)

Data(6).Plot = P75;
Data(6).xlabel = P75(1).Plot.XLabel;
Data(6).ylabel = P75(1).Plot.YLabel;
Data(6).xlim = 0.01;
Data(6).title = 'Vendor: White-box vs Black-box with fleet failures';

[fig_num,subplot_num] = plotting_subplots(40,Data,1,3,2);

clear Data;

% Now move the PFM of the conditional probability of failure om OC1 - OC5
% from cond_distr_params(X).PFM tp the respective plot_array_descr()
% structures.
for i=1:40,
    plot_array_descr(62+i-1).Plot.X = plot_array_descr(1).Plot.X;
    plot_array_descr(62+i-1).Plot.Y = cond_distr_params(5+i).pdf;
end

P76(1) = plot_array_descr(1);       % Conditional Prior (OC1)
P76(2) = plot_array_descr(62);      % AV1 Conditional Posterior (OC1)
P76(3) = plot_array_descr(67);      % AV2 Conditional Posterior (OC1) 
P76(4) = plot_array_descr(72);      % AV3 Conditional Posterior (OC1)
P76(5) = plot_array_descr(77);      % AV4 Conditional Posterior (OC1)
P76(6) = plot_array_descr(82);      % AV5 Conditional Posterior (OC1)
P76(7) = plot_array_descr(87);      % Vendor Conditional Posterior (OC1)

Data(1).Plot = P76;
Data(1).xlabel = P76(1).Plot.XLabel;
Data(1).ylabel = P76(1).Plot.YLabel;
Data(1).xlim = 0.02;
Data(1).title = 'OC1: Prior vs. Posterior PFM (no failures)';

P77(1) = plot_array_descr(2);       % Conditional Prior (OC1)
P77(2) = plot_array_descr(63);      % AV1 Conditional Posterior (OC2)
P77(3) = plot_array_descr(68);      % AV2 Conditional Posterior (OC3) 
P77(4) = plot_array_descr(73);      % AV3 Conditional Posterior (OC3)
P77(5) = plot_array_descr(78);      % AV4 Conditional Posterior (OC3)
P77(6) = plot_array_descr(83);      % AV5 Conditional Posterior (OC3)
P77(7) = plot_array_descr(88);      % Vendor Conditional Posterior (OC3)

Data(2).Plot = P77;
Data(2).xlabel = P77(1).Plot.XLabel;
Data(2).ylabel = P77(1).Plot.YLabel;
Data(2).xlim = 0.01;
Data(2).title = 'OC2: Prior vs. Posterior PFM (no failures)';

P78(1) = plot_array_descr(3);       % Conditional Prior (OC3)
P78(2) = plot_array_descr(64);      % AV1 Conditional Posterior (OC3)
P78(3) = plot_array_descr(69);      % AV2 Conditional Posterior (OC3) 
P78(4) = plot_array_descr(74);      % AV3 Conditional Posterior (OC3)
P78(5) = plot_array_descr(79);      % AV4 Conditional Posterior (OC3)
P78(6) = plot_array_descr(84);      % AV5 Conditional Posterior (OC3)
P78(7) = plot_array_descr(89);      % Vendor Conditional Posterior (OC3)

Data(3).Plot = P78;
Data(3).xlabel = P78(1).Plot.XLabel;
Data(3).ylabel = P78(1).Plot.YLabel;
Data(3).xlim = 0.01;
Data(3).title = 'OC3: Prior vs. Posterior PFM (no failures)';

P79(1) = plot_array_descr(4);       % Conditional Prior (OC4)
P79(2) = plot_array_descr(65);      % AV1 Conditional Posterior (OC4)
P79(3) = plot_array_descr(70);      % AV2 Conditional Posterior (OC4) 
P79(4) = plot_array_descr(75);      % AV3 Conditional Posterior (OC4)
P79(5) = plot_array_descr(80);      % AV4 Conditional Posterior (OC4)
P79(6) = plot_array_descr(85);      % AV5 Conditional Posterior (OC4)
P79(7) = plot_array_descr(90);      % Vendor Conditional Posterior (OC4)

Data(4).Plot = P79;
Data(4).xlabel = P79(1).Plot.XLabel;
Data(4).ylabel = P79(1).Plot.YLabel;
Data(4).xlim = 0.01;
Data(4).title = 'OC4: Prior vs. Posterior PFM (no failures)';

P80(1) = plot_array_descr(5);       % Conditional Prior (OC5)
P80(2) = plot_array_descr(66);      % AV1 Conditional Posterior (OC5)
P80(3) = plot_array_descr(71);      % AV2 Conditional Posterior (OC5) 
P80(4) = plot_array_descr(76);      % AV3 Conditional Posterior (OC5)
P80(5) = plot_array_descr(81);      % AV4 Conditional Posterior (OC5)
P80(6) = plot_array_descr(86);      % AV5 Conditional Posterior (OC5)
P80(7) = plot_array_descr(91);      % Vendor Conditional Posterior (OC5)

Data(5).Plot = P80;
Data(5).xlabel = P80(1).Plot.XLabel;
Data(5).ylabel = P80(1).Plot.YLabel;
Data(5).xlim = 0.01;
Data(5).title = 'OC5: Prior vs. Posterior PFM (no failures)';

P81(1) = plot_array_descr(20);      % system Prior (bootstrapining)
P81(2) = plot_array_descr(23);      % AV1 Posterior system PFM 
P81(3) = plot_array_descr(28);      % AV2 Posterior system PFM  
P81(4) = plot_array_descr(33);      % AV3 Posterior system PFM 
P81(5) = plot_array_descr(38);      % AV4 Posterior system PFM 
P81(6) = plot_array_descr(43);      % AV5 Posterior system PFM 
P81(7) = plot_array_descr(47);      % Vendor WB posterior system PFM 

Data(6).Plot = P81;
Data(6).xlabel = P81(1).Plot.XLabel;
Data(6).ylabel = P81(1).Plot.YLabel;
Data(6).xlim = 0.01;
Data(6).title = 'Prior vs. WB posterior system PFM (no failures)';

[fig_num,subplot_num] = plotting_subplots(50,Data,1,3,2);

clear Data;

% Now visualise the outcome of the cases with failurs (in AV3)

P82(1) = plot_array_descr(1);       % Conditional Prior (OC1)
P82(2) = plot_array_descr(62);      % AV1 Conditional Posterior (OC1)
P82(3) = plot_array_descr(67);      % AV2 Conditional Posterior (OC1) 
P82(4) = plot_array_descr(92);      % AV3 Conditional Posterior (OC1) with failures
P82(5) = plot_array_descr(77);      % AV4 Conditional Posterior (OC1)
P82(6) = plot_array_descr(82);      % AV5 Conditional Posterior (OC1)
P82(7) = plot_array_descr(97);      % Vendor Conditional Posterior (OC1) with failures

Data(1).Plot = P82;
Data(1).xlabel = P82(1).Plot.XLabel;
Data(1).ylabel = P82(1).Plot.YLabel;
Data(1).xlim = 0.02;
Data(1).title = 'OC1: Prior vs. Posterior PFM (with failures)';

P83(1) = plot_array_descr(2);       % Conditional Prior (OC1)
P83(2) = plot_array_descr(63);      % AV1 Conditional Posterior (OC2)
P83(3) = plot_array_descr(68);      % AV2 Conditional Posterior (OC3) 
P83(4) = plot_array_descr(93);      % AV3 Conditional Posterior (OC3) with failures
P83(5) = plot_array_descr(78);      % AV4 Conditional Posterior (OC3)
P83(6) = plot_array_descr(83);      % AV5 Conditional Posterior (OC3)
P83(7) = plot_array_descr(98);      % Vendor Conditional Posterior (OC3) with failures

Data(2).Plot = P83;
Data(2).xlabel = P83(1).Plot.XLabel;
Data(2).ylabel = P83(1).Plot.YLabel;
Data(2).xlim = 0.01;
Data(2).title = 'OC2: Prior vs. Posterior PFM (with failures)';

P84(1) = plot_array_descr(3);       % Conditional Prior (OC3)
P84(2) = plot_array_descr(64);      % AV1 Conditional Posterior (OC3)
P84(3) = plot_array_descr(69);      % AV2 Conditional Posterior (OC3) 
P84(4) = plot_array_descr(94);      % AV3 Conditional Posterior (OC3) with failures
P84(5) = plot_array_descr(79);      % AV4 Conditional Posterior (OC3)
P84(6) = plot_array_descr(84);      % AV5 Conditional Posterior (OC3)
P84(7) = plot_array_descr(99);      % Vendor Conditional Posterior (OC3) with failures

Data(3).Plot = P84;
Data(3).xlabel = P84(1).Plot.XLabel;
Data(3).ylabel = P84(1).Plot.YLabel;
Data(3).xlim = 0.01;
Data(3).title = 'OC3: Prior vs. Posterior PFM (with failures)';

P85(1) = plot_array_descr(4);       % Conditional Prior (OC4)
P85(2) = plot_array_descr(65);      % AV1 Conditional Posterior (OC4)
P85(3) = plot_array_descr(70);      % AV2 Conditional Posterior (OC4) 
P85(4) = plot_array_descr(95);      % AV3 Conditional Posterior (OC4) with failures
P85(5) = plot_array_descr(80);      % AV4 Conditional Posterior (OC4)
P85(6) = plot_array_descr(85);      % AV5 Conditional Posterior (OC4)
P85(7) = plot_array_descr(100);      % Vendor Conditional Posterior (OC4) with failures

Data(4).Plot = P85;
Data(4).xlabel = P85(1).Plot.XLabel;
Data(4).ylabel = P85(1).Plot.YLabel;
Data(4).xlim = 0.01;
Data(4).title = 'OC4: Prior vs. Posterior PFM (with failures)';

P86(1) = plot_array_descr(5);       % Conditional Prior (OC5)
P86(2) = plot_array_descr(66);      % AV1 Conditional Posterior (OC5)
P86(3) = plot_array_descr(71);      % AV2 Conditional Posterior (OC5) 
P86(4) = plot_array_descr(96);      % AV3 Conditional Posterior (OC5) with failures
P86(5) = plot_array_descr(81);      % AV4 Conditional Posterior (OC5)
P86(6) = plot_array_descr(86);      % AV5 Conditional Posterior (OC5)
P86(7) = plot_array_descr(101);      % Vendor Conditional Posterior (OC5) with failures

Data(5).Plot = P86;
Data(5).xlabel = P86(1).Plot.XLabel;
Data(5).ylabel = P86(1).Plot.YLabel;
Data(5).xlim = 0.01;
Data(5).title = 'OC5: Prior vs. Posterior PFM (with failures)';

P87(1) = plot_array_descr(20);      % system Prior (bootstrapining)
P87(2) = plot_array_descr(23);      % AV1 Posterior system PFM 
P87(3) = plot_array_descr(28);      % AV2 Posterior system PFM  
P87(4) = plot_array_descr(54);      % AV3 Posterior system PFM (with own failures)
P87(5) = plot_array_descr(38);      % AV4 Posterior system PFM 
P87(6) = plot_array_descr(43);      % AV5 Posterior system PFM 
P87(7) = plot_array_descr(61);      % Vendor WB posterior system PFM (with failures)

Data(6).Plot = P87;
Data(6).xlabel = P87(1).Plot.XLabel;
Data(6).ylabel = P87(1).Plot.YLabel;
Data(6).xlim = 0.01;
Data(6).title = 'Prior vs. WB posterior system PFM (with failures)';

[fig_num,subplot_num] = plotting_subplots(60,Data,1,3,2);

clear Data;

% Now plot only the prior distribution: conditional probablities of failure and the probability of system failure 

P90(1) = plot_array_descr(2);        % Distribution of Conditional probabilities of failure OC2
P90(2) = plot_array_descr(3);        % Distribution of Conditional probabilities of failure OC3
P90(3) = plot_array_descr(5);        % Distribution of Conditional probabilities of failure OC5

Data(1).Plot = P90;
Data(1).xlabel = P90(1).Plot.XLabel;
Data(1).ylabel = P90(1).Plot.YLabel;
Data(1).xlim = 0.01;
Data(1).title = 'Conditional Probability of failure (in OCs)';

P91(1) = plot_array_descr(1);        % % Distribution of Conditional probabilities of failure OC1
P91(2) = plot_array_descr(4);        % Distribution of Conditional probabilities of failure OC4

Data(2).Plot = P91;
Data(2).xlabel = P91(1).Plot.XLabel;
Data(2).ylabel = P91(1).Plot.YLabel;
Data(2).xlim = 0.02;
Data(2).title = 'Conditional Probability of failure (in OCs)';

P92(1) = plot_array_descr(20);        % % Distribution of Conditional probabilities of failure OC1

Data(3).Plot = P92;
Data(3).xlabel = P92(1).Plot.XLabel;
Data(3).ylabel = P92(1).Plot.YLabel;
Data(3).xlim = 0.01;
Data(3).title = 'Prior distribution of probability of system failure';

[fig_num,subplot_num] = plotting_subplots(70,Data,1,1,3);

clear Data;
P110(1) = plot_array_descr(20);      % Marginal Prior AV1
P110(2) = plot_array_descr(21);      % BB marginal posterior AV1 (own data)
P110(3) = plot_array_descr(23);      % WB marginal posterior AV1 (own data)
P110(4) = plot_array_descr(48);      % BB marginal posterior AV1 (fleet data, failures)
P110(5) = plot_array_descr(49);      % WB marginal posterior AV1 (fleet data, failures)

Data(1).Plot = P110;
Data(1).xlabel = P110(1).Plot.XLabel;
Data(1).ylabel = P110(1).Plot.YLabel;
Data(1).xlim = 0.01;
Data(1).title = 'AV1: White-box vs Black-box, fleet failures';

P111(1) = plot_array_descr(30);      % Marginal Prior AV3
P111(2) = plot_array_descr(52);      % BB marginal posterior AV3 (own data, failure)
P111(3) = plot_array_descr(53);      % WB marginal posterior AV3 (own data, failures)
P111(4) = plot_array_descr(54);      % BB marginal posterior AV3(fleet data, failure)
P111(5) = plot_array_descr(55);      % WB marginal posterior AV3(failures)

Data(2).Plot = P111;
Data(2).xlabel = P111(1).Plot.XLabel;
Data(2).ylabel = P111(1).Plot.YLabel;
Data(2).xlim = 0.01;
Data(2).title = 'AV3: White-box vs Black-box fleet failures';

P112(1) = plot_array_descr(45);      % Marginal Prior (bootstraping)
P112(2) = plot_array_descr(60);      % Vendor BB marginal posterior (fleet failures). 
P112(3) = plot_array_descr(61);      % Vendor WB marginal posterior (fleet failures)

Data(3).Plot = P112;
Data(3).xlabel = P112(1).Plot.XLabel;
Data(3).ylabel = P112(1).Plot.YLabel;
Data(3).xlim = 0.01;
Data(3).title = 'Vendor: White-box vs Black-box fleet failures';

[fig_num,subplot_num] = plotting_subplots(80,Data,1,1,3);

clear Data;

% Aggregated (vendor) data 
% Plot:
% - the distributions of the conditional probablities of failure (prior and
% posterior (with vendor data)); 
% - the marginal distributions of the probability of system failure: prior
% and BB/WB posterior (vendor data);

P95(1) = plot_array_descr(2);        % Prior Distribution of Conditional probabilities of failure OC2
P95(2) = plot_array_descr(3);        % Prior Distribution of Conditional probabilities of failure OC3
P95(3) = plot_array_descr(5);        % Prior Distribution of Conditional probabilities of failure OC5
P95(4) = plot_array_descr(88);        % Posterior Distribution of Conditional probabilities of failure OC2 (fleet data, no failure)
P95(5) = plot_array_descr(89);        % Posterior Distribution of Conditional probabilities of failure OC3 (fleet data, no failure)
P95(6) = plot_array_descr(91);        % Posterior Distribution of Conditional probabilities of failure OC5 (fleet data, no failure)

Data(1).Plot = P95;
Data(1).xlabel = P95(1).Plot.XLabel;
Data(1).ylabel = P95(1).Plot.YLabel;
Data(1).xlim = 0.01;
Data(1).title = 'Conditional Probability of failure (OC2, OC3, OC5)';

P96(1) = plot_array_descr(1);        % Prior Distribution of Conditional probabilities of failure OC1
P96(2) = plot_array_descr(4);        % Prior Distribution of Conditional probabilities of failure OC4
P96(3) = plot_array_descr(87);        % Posterior Distribution of Conditional probabilities of failure OC1 (fleet data, no failure)
P96(4) = plot_array_descr(90);        % Posterior Distribution of Conditional probabilities of failure OC4 (fleet data, no failure)

Data(2).Plot = P96;
Data(2).xlabel = P96(1).Plot.XLabel;
Data(2).ylabel = P96(1).Plot.YLabel;
Data(2).xlim = 0.02;
Data(2).title = 'Conditional Probability of failure (OC1, OC4)';

P97(1) = plot_array_descr(20);        % Prior Distribution of marginal probability of system failure
P97(2) = plot_array_descr(47);        % Posterior Distribution of marginal probability of system failure (WB, fleet data, no failure)
P97(3) = plot_array_descr(46);        % Posterior Distribution of marginal probability of system failure (BB, fleet data, no failure)

Data(3).Plot = P97;
Data(3).xlabel = P97(1).Plot.XLabel;
Data(3).ylabel = P97(1).Plot.YLabel;
Data(3).xlim = 0.01;
Data(3).title = 'Distribution of Probability of system failure';

P98(1) = plot_array_descr(2);        % Prior Distribution of Conditional probabilities of failure OC2
P98(2) = plot_array_descr(3);        % Prior Distribution of Conditional probabilities of failure OC3
P98(3) = plot_array_descr(5);        % Prior Distribution of Conditional probabilities of failure OC5
P98(4) = plot_array_descr(98);        % Posterior Distribution of Conditional probabilities of failure OC2 (fleet data, failure)
P98(5) = plot_array_descr(99);        % Posterior Distribution of Conditional probabilities of failure OC3 (fleet data, failure)
P98(6) = plot_array_descr(101);        % Posterior Distribution of Conditional probabilities of failure OC5 (fleet data, failure)

Data(4).Plot = P98;
Data(4).xlabel = P98(1).Plot.XLabel;
Data(4).ylabel = P98(1).Plot.YLabel;
Data(4).xlim = 0.01;
Data(4).title = 'Conditional Probability of failure (OC2, OC3, OC5): ';

P99(1) = plot_array_descr(1);        % Prior Distribution of Conditional probabilities of failure OC1
P99(2) = plot_array_descr(4);        % Prior Distribution of Conditional probabilities of failure OC4
P99(3) = plot_array_descr(97);        % Posterior Distribution of Conditional probabilities of failure OC1 (fleet data, no failure)
P99(4) = plot_array_descr(100);        % Posterior Distribution of Conditional probabilities of failure OC4 (fleet data, no failure)

Data(5).Plot = P99;
Data(5).xlabel = P99(1).Plot.XLabel;
Data(5).ylabel = P99(1).Plot.YLabel;
Data(5).xlim = 0.02;
Data(5).title = 'Conditional Probability of failure (OC1, OC4)';

P100(1) = plot_array_descr(20);        % Prior Distribution of marginal probability of system failure
P100(2) = plot_array_descr(61);        % Posterior Distribution of marginal probability of system failure (WB, fleet data, failure)
P100(3) = plot_array_descr(60);        % Posterior Distribution of marginal probability of system failure (BB, fleet data, failure)

Data(6).Plot = P100;
Data(6).xlabel = P100(1).Plot.XLabel;
Data(6).ylabel = P100(1).Plot.YLabel;
Data(6).xlim = 0.01;
Data(6).title = 'Distribution of Probability of system failure';

[fig_num,subplot_num] = plotting_subplots(90,Data,1,2,3);

clear Data;