Supplementary Material for
On the Evaluation of Unsupervised Outlier Detection: Measures, Datasets, and an Empirical Study
by G. O. Campos, A. Zimek, J. Sander, R. J. G. B. Campello, B. Micenková, E. Schubert, I. Assent and M. E. Houle
Data Mining and Knowledge Discovery 30(4): 891-927, 2016, DOI: 10.1007/s10618-015-0444-8

Parkinson (10% of outliers version#08)

The data set consists of medical data distinguishing healthy people from those suffering from Parkinson's disease. The latter were labeled as outliers.

Download all data set variants used (278.6 kB). You can also access the original data. (parkinsons.data)

Normalized, without duplicates

This version contains 22 attributes, 53 objects, 5 outliers (9.43%)

Download raw algorithm results (232.1 kB) Download raw algorithm evaluation table (10.4 kB)

Best Parameters

The following table contains the best (overall and per-method) results for each method and evaluation measure (when the same score was achieved twice, only the smallest k is given).
The Maximum F1-Measure is complimentary in addition to the measures in the original publication.

Algorithm k P@n Adj. P@n AP Adj. AP Max-F1 Adj. MF1 ROC AUC
KNN 43 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000
KNNW 3 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000
LOF 2 0.80000 0.77917 0.57905 0.53520 0.80000 0.77917 0.88750
LOF 29 0.80000 0.77917 0.86667 0.85278 0.88889 0.87731 0.95833
LOF 37 0.80000 0.77917 0.92500 0.91719 0.88889 0.87731 0.98750
SimplifiedLOF 6 0.80000 0.77917 0.76833 0.74420 0.80000 0.77917 0.97500
SimplifiedLOF 11 0.80000 0.77917 0.90000 0.88958 0.88889 0.87731 0.97917
SimplifiedLOF 13 0.80000 0.77917 0.94286 0.93690 0.88889 0.87731 0.99167
LoOP 7 0.80000 0.77917 0.75444 0.72887 0.80000 0.77917 0.97083
LoOP 14 0.80000 0.77917 0.89091 0.87955 0.88889 0.87731 0.97500
LoOP 46 0.80000 0.77917 0.90000 0.88958 0.88889 0.87731 0.97917
LDOF 19 0.80000 0.77917 0.84333 0.82701 0.80000 0.77917 0.96667
LDOF 20 0.80000 0.77917 0.87692 0.86410 0.88889 0.87731 0.96667
LDOF 51 0.80000 0.77917 0.90000 0.88958 0.88889 0.87731 0.97917
ODIN 29 0.80000 0.77917 0.75167 0.72580 0.80000 0.77917 0.91458
ODIN 43 0.80000 0.77917 0.92667 0.91903 0.90909 0.89962 0.99167
ODIN 45 0.80000 0.77917 0.93333 0.92639 0.90909 0.89962 0.99167
FastABOD 6 0.80000 0.77917 0.88500 0.87302 0.80000 0.77917 0.98333
FastABOD 8 0.80000 0.77917 0.94286 0.93690 0.88889 0.87731 0.99167
KDEOS 19 0.60000 0.55833 0.54500 0.49760 0.76923 0.74519 0.95417
KDEOS 21 0.60000 0.55833 0.65952 0.62406 0.83333 0.81597 0.97083
KDEOS 52 0.60000 0.55833 0.84444 0.82824 0.75000 0.72396 0.97500
LDF 23 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000
INFLO 39 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000
COF 7 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000 1.00000

Plots

Precision at n
Adjusted precision at n
Average precision
Adjusted average precision
Maximum F1 score
Adjusted maximum F1 score
ROC AUC
Diversity
A: KNN, B: KNNW, C: LOF, D: SimplifiedLOF, E: LoOP, F: LDOF
G: ODIN, H: KDEOS, I: COF, J: FastABOD, K: LDF, L: INFLO

Not normalized, without duplicates

This version contains 22 attributes, 53 objects, 5 outliers (9.43%)

Download raw algorithm results (234.8 kB) Download raw algorithm evaluation table (18.9 kB)

Best Parameters

The following table contains the best (overall and per-method) results for each method and evaluation measure (when the same score was achieved twice, only the smallest k is given).
The Maximum F1-Measure is complimentary in addition to the measures in the original publication.

Algorithm k P@n Adj. P@n AP Adj. AP Max-F1 Adj. MF1 ROC AUC
KNN 1 0.20000 0.11667 0.27612 0.20072 0.50000 0.44792 0.80000
KNN 2 0.20000 0.11667 0.29541 0.22202 0.50000 0.44792 0.80625
KNNW 1 0.40000 0.33750 0.27602 0.20061 0.44444 0.38657 0.80000
KNNW 2 0.00000 -0.10417 0.26940 0.19330 0.53333 0.48472 0.80000
KNNW 3 0.20000 0.11667 0.29292 0.21926 0.53333 0.48472 0.80833
KNNW 6 0.20000 0.11667 0.29541 0.22202 0.50000 0.44792 0.80417
LOF 2 0.40000 0.33750 0.39336 0.33016 0.40000 0.33750 0.76250
LOF 3 0.20000 0.11667 0.47279 0.41787 0.61538 0.57532 0.76667
LOF 4 0.40000 0.33750 0.35802 0.29115 0.61538 0.57532 0.80417
SimplifiedLOF 3 0.20000 0.11667 0.39586 0.33293 0.61538 0.57532 0.87083
SimplifiedLOF 4 0.20000 0.11667 0.49586 0.44334 0.61538 0.57532 0.87500
SimplifiedLOF 5 0.40000 0.33750 0.38238 0.31805 0.61538 0.57532 0.88333
LoOP 3 0.40000 0.33750 0.34824 0.28035 0.44444 0.38657 0.81667
LoOP 5 0.40000 0.33750 0.48763 0.43426 0.53333 0.48472 0.88333
LDOF 5 0.40000 0.33750 0.38539 0.32137 0.40000 0.33750 0.70000
LDOF 6 0.40000 0.33750 0.44394 0.38602 0.47059 0.41544 0.76250
ODIN 5 0.16000 0.07250 0.23333 0.15347 0.40000 0.33750 0.81042
ODIN 6 0.25000 0.17188 0.22703 0.14651 0.38095 0.31647 0.78542
FastABOD 3 0.20000 0.11667 0.22295 0.14200 0.40000 0.33750 0.73750
FastABOD 4 0.20000 0.11667 0.24851 0.17023 0.44444 0.38657 0.77500
KDEOS 7 0.40000 0.33750 0.28673 0.21244 0.40000 0.33750 0.75417
KDEOS 8 0.40000 0.33750 0.31085 0.23906 0.44444 0.38657 0.81250
KDEOS 24 0.40000 0.33750 0.40632 0.34447 0.53333 0.48472 0.80417
KDEOS 26 0.20000 0.11667 0.28667 0.21237 0.57143 0.52679 0.77917
LDF 2 0.20000 0.11667 0.47941 0.42518 0.61538 0.57532 0.81667
LDF 3 0.60000 0.55833 0.40750 0.34578 0.61538 0.57532 0.77917
LDF 4 0.60000 0.55833 0.44500 0.38719 0.72727 0.69886 0.82083
INFLO 2 0.40000 0.33750 0.39426 0.33116 0.44444 0.38657 0.72500
INFLO 3 0.20000 0.11667 0.47817 0.42382 0.57143 0.52679 0.87500
COF 3 0.40000 0.33750 0.35821 0.29135 0.54545 0.49811 0.86667
COF 4 0.40000 0.33750 0.50556 0.45405 0.57143 0.52679 0.87083
COF 6 0.40000 0.33750 0.44070 0.38244 0.57143 0.52679 0.87500

Plots

Precision at n
Adjusted precision at n
Average precision
Adjusted average precision
Maximum F1 score
Adjusted maximum F1 score
ROC AUC
Diversity
A: KNN, B: KNNW, C: LOF, D: SimplifiedLOF, E: LoOP, F: LDOF
G: ODIN, H: KDEOS, I: COF, J: FastABOD, K: LDF, L: INFLO