function results_evolution_strategies
%RESULTS_CONTROL_STRESS_ANALYSIS Plot distribution of strategies for each
%trial
addpath(fullfile(fileparts(mfilename('fullpath')), '../extern/export_fig'));
addpath(fullfile(fileparts(mfilename('fullpath')), '../'));
% global data initialized elsewhere
global g_trajectories_trial;
global g_segments_classification;
global g_long_trajectories_idx;
global g_partitions;
trajectories_latency = arrayfun( @(t) t.compute_feature(g_config.FEATURE_LATENCY), g_trajectories.items);
% classify trajectories
cache_trajectories_classification;
distr = g_segments_classification.classes_distribution(g_partitions(g_long_trajectories_idx));
% plot distribution for each trial
data = [];
for t = 1:g_config.TRIALS
data = [data, ...
arrayfun( @(c) sum(distr(g_trajectories_trial(g_long_trajectories_idx) == t, c)), ...
1:g_segments_classification.nclasses...
)'];
end
% normalize the data
data = 100*data ./ repmat(sum(data), size(data, 1), 1);
figure(321);
bar(data', 'Stack');
colormap(g_config.CLASSES_COLORMAP);
xlabel('trial', 'FontSize', g_config.FONT_SIZE);
ylabel('percentage', 'FontSize', g_config.FONT_SIZE);
box off;
export_fig(fullfile(g_config.OUTPUT_DIR, 'distribution_strat_trials.eps'));
% do the same for very long trajectories (latency >80 seconds)
data = [];
for t = 1:g_config.TRIALS
data = [data, ...
arrayfun( @(c) sum(distr(g_trajectories_trial(g_long_trajectories_idx) == t ...
& trajectories_latency(g_long_trajectories_idx) > 80, c)), ...
1:g_segments_classification.nclasses...
)'];
end
figure;
data = 100*data ./ repmat(sum(data), size(data, 1), 1);
bar(data', 'Stack');
colormap(g_config.CLASSES_COLORMAP);
xlabel('trial', 'FontSize', g_config.FONT_SIZE);
ylabel('percentage', 'FontSize', g_config.FONT_SIZE);
box off;
export_fig(fullfile(g_config.OUTPUT_DIR, 'distribution_strat_trials_80.eps'));
end