6
$\begingroup$

I would like to compare outlier detection algorithms. I am not sure if area under roc or under precision recall curve is the measure to use.

A quick test in matlab gives me strange results. I try to get the ROC and PR values for a perfect classification:

% true labels
outlier = 1;
normal  = 0;
% 99% normal data 1% outlier
label = normal*ones(1000,1); 
label(1:10) = outlier;  

% scores of the algorithm
% assume the prediction is perfect
score = label;

[~,~,~,AUC] = perfcurve(label,score,outlier)  % AUC = 1 
[~,~,~,PR] = perfcurve(label,score,outlier, 'xCrit', 'reca', 'yCrit', 'prec') % PR = 0

Why is the area under PR = 0? Shouldn't it be 1?

The AUC ROC also gives the expected result if evaluated with a constant classifier which is 0.5:

label = real( rand(1000,1) > 0.99 );     % 99% normal data 1% outlier
score = zeros(1000,1);                   % always predicting zero
[~,~,~,AUC] = perfcurve(label,score,1)   % AUC = 0.5

So why is the area under PR curve considered better to compare outlier detection algorithms?

Edit: I do not aim for a matlab specific answer to my question. This is just a MWE for you.

I just wonder why so many publications recommend the AUC PR to be better suited than ROC for imbalanced datasets as their are typical for outlier detection. My quick proof of concept above seems to indicate the opposite.

$\endgroup$
  • $\begingroup$ Could you specify what language your code is in? Is it MATLAB? $\endgroup$ – Alex R. Jun 3 '16 at 19:34
  • $\begingroup$ Yes it is MATLAB $\endgroup$ – Manuel Schmidt Jun 4 '16 at 10:15
4
+50
$\begingroup$

The problem is with your example that it is possible to have zero $tp$ and zero $fp$, therefore the precision $prec = tp/(tp+fp)$ becomes undefined because we divide by zero. Because of this the PR curve only contains points for one $x$-value, and therefore the area under the PR curve becomes zero for your example.

You can see this by plotting the PR curve:

[X,Y,T,PR] = perfcurve(label,score,1, 'xCrit', 'reca', 'yCrit', 'prec') % PR = 0
figure
scatter(X,Y)
xlabel('recall')
ylabel('precision')

So plotting a PR curve doesn't really work well when all your scores are the same.

To gain more insights between the difference of the PR curve and the ROC curve, compare these two prediction lists. We consider the case where we predict all zeros, and predict one 1, but it should be zero (score1). This one doesnt work very well, it predicts 0 everywhere, except for one object where it predicts 1 where it should be zero. We consider another case, where we predict one 1 correctly, and the rest we classify as 0. Here we thus predict 1 one correctly, and the rest we classify as 0. We compare the area under the PR curve and the area under the ROC.

outlier = 1;
normal  = 0;
% 99% normal data 1% outlier
label = normal*ones(1000,1); 
label(1:10) = outlier;  

%label = real( rand(1000,1) > 0.99 );     % 99% normal data 1% outlier
score1 = [zeros(999,1);1]; % predict everything as zero, and one mistake 
score2 = [1;zeros(999,1)]; % predict everything as zero, and one 1 correct 

[X,Y,T,AUC1] = perfcurve(label,score1,1)
% AUC1 = 0.5
[X,Y,T,AUC2] = perfcurve(label,score2,1)
% AUC2 = 0.55

[X,Y,T,PR1] = perfcurve(label,score1,1, 'xCrit', 'reca', 'yCrit', 'prec') 
% PR1 = 0.005 
[X,Y,T,PR2] = perfcurve(label,score2,1, 'xCrit', 'reca', 'yCrit', 'prec') 
% PR2 = 0.4545

Observe that the AUC varies little between score1 and score2. However, the area under the PR curve is significantly different. It rewards score2 much more than score1. This indicates it is better suited to outlier detection: it rewards detecting the outlier much more than the AUC. In case of outlier detection you would prefer score2 much more, since it predicts the 1 that you want to detect correctly, while score1 predicts a 1 for a zero and never catches any outliers.

In general, the AUC is more informative to give an idea how well your predictions work for varying priors. Thus the AUC characterizes how well the classifier works for varying number of ones and zeros.

The PR curves indicates more well how it performs for the current class imbalance considered. Therefore the PR curve is more interesting for you: it takes into account there are little 1's in your dataset than 0's. Because you are only interested in this case when you are interested in outlier detection, the PR curve is more informative.

While the AUC characterizes how your predictions would do if there are much more 1's as well.

For more information see also:

https://www.quora.com/What-is-Precision-Recall-PR-curve

ROC vs precision-and-recall curves

Finally, you might be interested in how to compute an ROC / PR curve, a detailed explanation is given here for ROC curves:

http://blogs.sas.com/content/iml/2011/07/29/computing-an-roc-curve-from-basic-principles.html

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.