load dsSmall.mat
[cPrm, logLike1, rr1]=nbcTrain(ds);
[computedClass, logLike2, rr2, hitIndex]=nbcEval(dsTest, cPrm);
logLike=diff(logLike2);
id01=find(dsTest.output==1);
id02=find(dsTest.output==2);
data01=logLike(id01);
data02=logLike(id02);
opt=detPlot('defaultOpt');
opt.scaleMode='linear';
detPlot(data01, data02, opt);
opt=detGet('defaultOpt');
figure; [th, fp, fn]=detGet(data01, data02, opt, 1);
%plot(dsTest.output, logLike, 'o');
%%
figure;
costPair=[1 1];
fnCosts=1:6;
for i=1:length(fnCosts)
	costPair(2)=fnCosts(i);
	[costMin(i), thBest(i), costVec, thVec, misclassifiedCount{i}]=cost4binClass(dsTest.output, logLike, costPair, 1); hold on
end
hold off
%%
figure;
plot(fnCosts, thBest, 'o-');
for i=1:length(fnCosts)
	h=text(fnCosts(i), thBest(i), sprintf('  %s', mat2str(misclassifiedCount{i})));
	set(h, 'rotation', 30);
end
xlabel('FN costs'); ylabel('Thresholds'); title('Best thresholds vs. FN costs');
%% Lift chart
figure;
gt=dsTest.output;
binCount=10;
showPlot=1;
[binPercent, binSize]=liftChart(logLike, gt, binCount, showPlot);