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

Not normalized data with duplicates, Datasets with less or equal 5% outliers

Average performance across all data sets, per method.
Best results were always chosen by ROC AUC performance.

Adjusted Average Precision

All k
Best k
Best k±5
Best k±10

Average Precision

All k
Best k
Best k±5
Best k±10

Adjusted Precision@n

All k
Best k
Best k±5
Best k±10

Precision@n

All k
Best k
Best k±5
Best k±10

ROC AUC

All k
Best k
Best k±5
Best k±10

Adjusted Maximum F1

All k
Best k
Best k±5
Best k±10

Maximum F1

All k
Best k
Best k±5
Best k±10