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

KDDCup99

This data set captures different types of network intrusions or attacks. We follow the preprocessing performed by Nguyen and Gopalkrishnan [1,2], without removing the categorical attributes, using U2R outlier and normal as inlier. U2R as referred to by Nguyen and Gopalkrishnan contains the following attacks: buffer_overflow, ftp_write, imap, load_module, multihop, nmap, perl, phf, pod, rootkit, and teardrop. These attacks have 246 instances of the database, defined as outliers. After the preprocessing, this database has 38 numerical attributes, 3 categorical attributes and 60,839 instances, divided into 60,593 inliers and 246 outliers. This dataset contains 12,726 duplicates (46 outliers and 12,680 inliers). A slightly different version of this data set was also used in [3,4].

It was not feasible to run FastABOD for this data set due to its size and therefore the results for this method are missing. For the version with duplicates, also generating results of LDOF was computationally prohibitive.

References

[1] H. V. Nguyen and V. Gopalkrishnan. Feature extraction for outlier detection in high-dimensional spaces. J. Mach. Learn. Res., Proceedings Track 10:66-75, 2010.
[2] H. V. Nguyen, H. H. Ang, and V. Gopalkrishnan. Mining outliers with ensemble of heterogeneous detectors on random subspaces. In Proc. DASFAA, pages 368-383, 2010.
[3] H.-P. Kriegel, P. Kroeger, E. Schubert, and A. Zimek. Interpreting and unifying outlier scores. In Proc. SDM, pages 13-24, 2011
[4] E. Schubert, R. Wojdanowski, A. Zimek, and H.-P. Kriegel. On evaluation of outlier rankings and outlier scores. In Proc. SDM, pages 1047-1058, 2012.

Download all data set variants used (15.7 MB). You can also access the original data. (corrected.gz)

Normalized, without duplicates, idf weighted categorial attributes

This version contains 40 attributes, 48113 objects, 200 outliers (0.42%)

Download raw algorithm results (333.2 MB) Download raw algorithm evaluation table (62.0 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 80 0.43500 0.43264 0.25543 0.25232 0.47748 0.47530 0.98910
KNN 89 0.43500 0.43264 0.54723 0.54534 0.56140 0.55957 0.98974
KNN 92 0.43500 0.43264 0.54939 0.54751 0.56940 0.56760 0.98967
KNN 94 0.43500 0.43264 0.54901 0.54713 0.57143 0.56964 0.98963
KNNW 19 0.21500 0.21172 0.08512 0.08130 0.22407 0.22083 0.91802
KNNW 100 0.17500 0.17156 0.14322 0.13964 0.29428 0.29133 0.98541
LOF 1 0.01500 0.01089 0.00585 0.00170 0.02215 0.01807 0.51514
LOF 100 0.00000 -0.00417 0.01275 0.00862 0.03463 0.03060 0.84894
SimplifiedLOF 1 0.00000 -0.00417 0.00484 0.00068 0.01392 0.00981 0.52548
SimplifiedLOF 37 0.00000 -0.00417 0.00810 0.00396 0.03056 0.02651 0.60359
SimplifiedLOF 62 0.00000 -0.00417 0.00866 0.00453 0.02565 0.02158 0.66796
LoOP 5 0.02500 0.02093 0.00735 0.00321 0.02867 0.02462 0.56709
LoOP 25 0.01500 0.01089 0.00926 0.00513 0.04198 0.03798 0.60272
LoOP 62 0.00000 -0.00417 0.01126 0.00713 0.03305 0.02901 0.70088
LoOP 65 0.00000 -0.00417 0.01124 0.00712 0.03372 0.02968 0.70313
ODIN 64 0.04138 0.03738 0.01452 0.01040 0.05361 0.04966 0.67052
ODIN 77 0.04433 0.04034 0.01463 0.01051 0.04955 0.04558 0.69974
ODIN 100 0.03500 0.03097 0.01757 0.01347 0.04789 0.04392 0.80767
KDEOS 37 0.05000 0.04603 0.00875 0.00462 0.05038 0.04641 0.54202
KDEOS 68 0.03000 0.02595 0.00841 0.00427 0.05128 0.04732 0.60510
KDEOS 70 0.04000 0.03599 0.00849 0.00435 0.05674 0.05280 0.60321
KDEOS 73 0.05500 0.05106 0.00845 0.00431 0.05542 0.05147 0.60013
LDF 4 0.01754 0.01344 0.00655 0.00240 0.02525 0.02118 0.55452
LDF 90 0.00000 -0.00417 0.02304 0.01896 0.09192 0.08813 0.87696
INFLO 1 0.01000 0.00587 0.00586 0.00171 0.02471 0.02064 0.52189
INFLO 11 0.00000 -0.00417 0.00729 0.00314 0.03302 0.02899 0.58058
INFLO 56 0.00000 -0.00417 0.00966 0.00552 0.02913 0.02507 0.70334
COF 5 0.02500 0.02093 0.00691 0.00277 0.02801 0.02395 0.54261
COF 58 0.00000 -0.00417 0.00971 0.00558 0.03712 0.03310 0.66602
COF 67 0.00000 -0.00417 0.00977 0.00563 0.03546 0.03144 0.67006

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

Normalized, without duplicates, 1-of-n encoding

This version contains 79 attributes, 48113 objects, 200 outliers (0.42%)

Download raw algorithm results (330.4 MB) Download raw algorithm evaluation table (63.3 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 83 0.44000 0.43766 0.36243 0.35976 0.51757 0.51556 0.99030
KNN 84 0.44000 0.43766 0.47318 0.47098 0.54671 0.54482 0.99044
KNN 86 0.44000 0.43766 0.51395 0.51192 0.56738 0.56557 0.99041
KNN 90 0.44000 0.43766 0.53886 0.53693 0.56738 0.56557 0.99034
KNNW 4 0.21000 0.20670 0.08389 0.08006 0.22222 0.21898 0.84409
KNNW 100 0.13500 0.13139 0.13046 0.12683 0.26623 0.26317 0.98538
LOF 1 0.02500 0.02093 0.00632 0.00217 0.03488 0.03086 0.51100
LOF 15 0.00000 -0.00417 0.01037 0.00624 0.05439 0.05045 0.59062
LOF 100 0.00000 -0.00417 0.01117 0.00704 0.03319 0.02915 0.81301
SimplifiedLOF 6 0.01500 0.01089 0.00971 0.00557 0.04348 0.03949 0.59506
SimplifiedLOF 18 0.00000 -0.00417 0.01049 0.00636 0.04981 0.04584 0.61206
SimplifiedLOF 26 0.00000 -0.00417 0.01060 0.00647 0.04734 0.04337 0.60295
SimplifiedLOF 53 0.00000 -0.00417 0.01045 0.00632 0.03130 0.02725 0.69030
LoOP 5 0.04000 0.03599 0.00859 0.00446 0.04420 0.04021 0.56714
LoOP 26 0.03000 0.02595 0.01341 0.00929 0.05977 0.05585 0.62965
LoOP 64 0.00000 -0.00417 0.01174 0.00761 0.03522 0.03119 0.70759
ODIN 50 0.03732 0.03330 0.01436 0.01024 0.06320 0.05929 0.68842
ODIN 61 0.04091 0.03691 0.01473 0.01062 0.06034 0.05642 0.67660
ODIN 63 0.03828 0.03427 0.01490 0.01079 0.06100 0.05708 0.67395
ODIN 100 0.03645 0.03243 0.01350 0.00938 0.04291 0.03892 0.78560
KDEOS 36 0.05000 0.04603 0.01168 0.00756 0.05155 0.04759 0.56170
KDEOS 37 0.05000 0.04603 0.01209 0.00796 0.05319 0.04924 0.56531
KDEOS 43 0.05000 0.04603 0.01022 0.00609 0.06116 0.05724 0.59333
KDEOS 69 0.04000 0.03599 0.00940 0.00527 0.05505 0.05110 0.63055
LDF 82 0.03635 0.03233 0.02612 0.02206 0.06795 0.06406 0.86298
LDF 96 0.03679 0.03277 0.02661 0.02255 0.06849 0.06460 0.85326
LDF 97 0.03631 0.03229 0.02679 0.02273 0.06780 0.06391 0.85473
INFLO 1 0.02000 0.01591 0.00617 0.00202 0.03422 0.03019 0.50999
INFLO 23 0.00000 -0.00417 0.01138 0.00725 0.05530 0.05136 0.64183
INFLO 24 0.00000 -0.00417 0.01140 0.00727 0.05458 0.05063 0.64427
INFLO 50 0.00000 -0.00417 0.00947 0.00533 0.03256 0.02852 0.67834
COF 5 0.03500 0.03097 0.00820 0.00406 0.04107 0.03706 0.55380
COF 25 0.02500 0.02093 0.01195 0.00782 0.05344 0.04948 0.56228
COF 72 0.00000 -0.00417 0.01137 0.00725 0.05618 0.05224 0.66425
COF 75 0.00000 -0.00417 0.01119 0.00706 0.05807 0.05414 0.64819

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

Normalized, without duplicates, categorial attributes removed

This version contains 40 attributes, 48113 objects, 200 outliers (0.42%)

Download raw algorithm results (333.2 MB) Download raw algorithm evaluation table (62.1 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 80 0.43500 0.43264 0.25543 0.25232 0.47748 0.47530 0.98910
KNN 89 0.43500 0.43264 0.54723 0.54534 0.56140 0.55957 0.98974
KNN 92 0.43500 0.43264 0.54939 0.54751 0.56940 0.56760 0.98967
KNN 94 0.43500 0.43264 0.54901 0.54713 0.57143 0.56964 0.98963
KNNW 19 0.21500 0.21172 0.08512 0.08130 0.22407 0.22083 0.91802
KNNW 100 0.17500 0.17156 0.14322 0.13964 0.29428 0.29133 0.98541
LOF 1 0.01500 0.01089 0.00585 0.00170 0.02215 0.01807 0.51514
LOF 100 0.00000 -0.00417 0.01275 0.00862 0.03463 0.03060 0.84894
SimplifiedLOF 1 0.00000 -0.00417 0.00484 0.00068 0.01392 0.00981 0.52548
SimplifiedLOF 37 0.00000 -0.00417 0.00810 0.00396 0.03056 0.02651 0.60359
SimplifiedLOF 62 0.00000 -0.00417 0.00866 0.00453 0.02565 0.02158 0.66796
LoOP 5 0.02500 0.02093 0.00735 0.00321 0.02867 0.02462 0.56709
LoOP 25 0.01500 0.01089 0.00926 0.00513 0.04198 0.03798 0.60272
LoOP 62 0.00000 -0.00417 0.01126 0.00713 0.03305 0.02901 0.70088
LoOP 65 0.00000 -0.00417 0.01124 0.00712 0.03372 0.02968 0.70313
ODIN 64 0.04138 0.03738 0.01452 0.01040 0.05361 0.04966 0.67052
ODIN 77 0.04433 0.04034 0.01463 0.01051 0.04955 0.04558 0.69974
ODIN 100 0.03500 0.03097 0.01757 0.01347 0.04789 0.04392 0.80767
KDEOS 37 0.05000 0.04603 0.00875 0.00462 0.05038 0.04641 0.54202
KDEOS 68 0.03000 0.02595 0.00841 0.00427 0.05128 0.04732 0.60510
KDEOS 70 0.04000 0.03599 0.00849 0.00435 0.05674 0.05280 0.60321
KDEOS 73 0.05500 0.05106 0.00845 0.00431 0.05542 0.05147 0.60013
LDF 4 0.01754 0.01344 0.00655 0.00240 0.02525 0.02118 0.55452
LDF 90 0.00000 -0.00417 0.02304 0.01896 0.09192 0.08813 0.87696
INFLO 1 0.01000 0.00587 0.00586 0.00171 0.02471 0.02064 0.52189
INFLO 11 0.00000 -0.00417 0.00729 0.00314 0.03302 0.02899 0.58058
INFLO 56 0.00000 -0.00417 0.00966 0.00552 0.02913 0.02507 0.70334
COF 5 0.02500 0.02093 0.00691 0.00277 0.02801 0.02395 0.54261
COF 58 0.00000 -0.00417 0.00971 0.00558 0.03712 0.03310 0.66602
COF 67 0.00000 -0.00417 0.00977 0.00563 0.03546 0.03144 0.67006

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

Normalized, duplicates, idf weighted categorial attributes

This version contains 41 attributes, 60839 objects, 246 outliers (0.40%)

Download raw algorithm results (349.2 MB) Download raw algorithm evaluation table (63.1 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 83 0.41463 0.41226 0.28655 0.28366 0.44712 0.44487 0.99022
KNN 88 0.40650 0.40409 0.49325 0.49119 0.50299 0.50098 0.99045
KNN 90 0.40244 0.40001 0.50616 0.50415 0.50755 0.50555 0.99039
KNN 91 0.40244 0.40001 0.50510 0.50309 0.50909 0.50710 0.99035
KNNW 16 0.21951 0.21634 0.07660 0.07285 0.22626 0.22312 0.84059
KNNW 100 0.19919 0.19594 0.13598 0.13247 0.23494 0.23183 0.98353
LOF 15 0.02439 0.02043 0.00956 0.00554 0.04635 0.04248 0.60209
LOF 100 0.00000 -0.00406 0.01197 0.00796 0.03204 0.02811 0.83750
SimplifiedLOF 2 0.03198 0.02805 0.00713 0.00310 0.03729 0.03338 0.53727
SimplifiedLOF 18 0.01917 0.01519 0.00946 0.00544 0.04063 0.03674 0.60936
SimplifiedLOF 84 0.00000 -0.00406 0.00719 0.00316 0.02607 0.02212 0.62948
LoOP 18 0.02846 0.02451 0.01080 0.00679 0.04888 0.04502 0.61388
LoOP 34 0.02033 0.01635 0.01278 0.00878 0.04918 0.04532 0.68076
LoOP 43 0.02033 0.01635 0.01284 0.00883 0.05594 0.05211 0.65229
LoOP 47 0.02033 0.01635 0.01287 0.00887 0.05377 0.04993 0.65246
ODIN 71 0.04878 0.04492 0.01571 0.01172 0.06378 0.05998 0.68528
ODIN 77 0.04388 0.04000 0.01573 0.01173 0.06676 0.06297 0.70657
ODIN 100 0.02846 0.02451 0.01767 0.01368 0.06172 0.05791 0.80962
KDEOS 40 0.04472 0.04084 0.01023 0.00621 0.05051 0.04665 0.59015
KDEOS 44 0.04472 0.04084 0.01060 0.00659 0.04785 0.04398 0.60182
KDEOS 50 0.02846 0.02451 0.00874 0.00471 0.03819 0.03428 0.61512
LDF 19 0.02378 0.01981 0.00853 0.00450 0.03589 0.03198 0.57009
LDF 93 0.00000 -0.00406 0.02181 0.01784 0.06433 0.06053 0.88800
LDF 94 0.00000 -0.00406 0.02139 0.01742 0.07033 0.06656 0.88543
INFLO 15 0.02439 0.02043 0.00923 0.00521 0.04418 0.04030 0.58808
INFLO 20 0.00000 -0.00406 0.00940 0.00538 0.04629 0.04242 0.61832
INFLO 22 0.00000 -0.00406 0.00976 0.00574 0.04590 0.04202 0.62589
INFLO 100 0.00000 -0.00406 0.00695 0.00292 0.02313 0.01916 0.64850
COF 5 0.02846 0.02451 0.00669 0.00266 0.02917 0.02523 0.54745
COF 22 0.02439 0.02043 0.00979 0.00577 0.04664 0.04277 0.59421
COF 35 0.01626 0.01227 0.01077 0.00675 0.03796 0.03405 0.62559

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

Normalized, duplicates, 1-of-n encoding

This version contains 79 attributes, 60839 objects, 246 outliers (0.40%)

Download raw algorithm results (346.3 MB) Download raw algorithm evaluation table (64.3 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 85 0.38618 0.38369 0.36786 0.36529 0.48159 0.47948 0.98982
KNN 86 0.39024 0.38777 0.36719 0.36462 0.48159 0.47948 0.98979
KNN 88 0.39024 0.38777 0.48118 0.47908 0.50909 0.50710 0.98972
KNN 90 0.39024 0.38777 0.48217 0.48006 0.50909 0.50710 0.98966
KNNW 8 0.20325 0.20002 0.07680 0.07305 0.20971 0.20650 0.81234
KNNW 19 0.17073 0.16736 0.08079 0.07706 0.23847 0.23538 0.87528
KNNW 100 0.15447 0.15104 0.11989 0.11632 0.23577 0.23267 0.98273
LOF 13 0.01626 0.01227 0.01043 0.00641 0.05441 0.05058 0.58504
LOF 16 0.01626 0.01227 0.01119 0.00718 0.05380 0.04996 0.62139
LOF 100 0.00000 -0.00406 0.01091 0.00689 0.03139 0.02745 0.82330
SimplifiedLOF 1 0.02662 0.02266 0.00640 0.00237 0.03627 0.03236 0.52355
SimplifiedLOF 16 0.01626 0.01227 0.01029 0.00627 0.04762 0.04375 0.60858
SimplifiedLOF 34 0.00000 -0.00406 0.00966 0.00564 0.04012 0.03622 0.62814
LoOP 20 0.03659 0.03267 0.01192 0.00791 0.05372 0.04988 0.63248
LoOP 34 0.02033 0.01635 0.01445 0.01045 0.06088 0.05706 0.68726
LoOP 41 0.01626 0.01227 0.01392 0.00992 0.06434 0.06054 0.65501
ODIN 53 0.03363 0.02971 0.01461 0.01061 0.06073 0.05692 0.64690
ODIN 61 0.04201 0.03812 0.01524 0.01124 0.05855 0.05473 0.65985
ODIN 62 0.04254 0.03865 0.01503 0.01104 0.05778 0.05396 0.65955
ODIN 100 0.02422 0.02026 0.01327 0.00926 0.04944 0.04558 0.78924
KDEOS 43 0.04878 0.04492 0.01244 0.00843 0.05728 0.05345 0.60307
KDEOS 50 0.03659 0.03267 0.01130 0.00728 0.05336 0.04952 0.61999
KDEOS 75 0.05285 0.04900 0.01049 0.00648 0.05941 0.05559 0.56306
KDEOS 76 0.05691 0.05308 0.01031 0.00630 0.05833 0.05451 0.56158
LDF 87 0.03692 0.03301 0.02674 0.02278 0.06834 0.06456 0.88959
INFLO 13 0.01626 0.01227 0.00970 0.00568 0.04954 0.04569 0.58077
INFLO 21 0.00000 -0.00406 0.01103 0.00702 0.05288 0.04904 0.64316
INFLO 25 0.00000 -0.00406 0.01095 0.00693 0.05044 0.04658 0.64848
INFLO 27 0.00000 -0.00406 0.01114 0.00712 0.05155 0.04770 0.64641
COF 5 0.02846 0.02451 0.00680 0.00276 0.03285 0.02892 0.54372
COF 35 0.00813 0.00410 0.01278 0.00877 0.06275 0.05894 0.61605
COF 36 0.00813 0.00410 0.01282 0.00881 0.06184 0.05803 0.61160
COF 60 0.00813 0.00410 0.01136 0.00735 0.06996 0.06619 0.53901

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

Normalized, duplicates, categorial attributes removed

This version contains 38 attributes, 60839 objects, 246 outliers (0.40%)

Download raw algorithm results (349.3 MB) Download raw algorithm evaluation table (63.9 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 83 0.40650 0.40409 0.29725 0.29440 0.45885 0.45666 0.98107
KNN 92 0.39024 0.38777 0.48047 0.47836 0.50755 0.50555 0.98304
KNN 95 0.38618 0.38369 0.48015 0.47803 0.50909 0.50710 0.98316
KNN 100 0.38618 0.38369 0.47647 0.47435 0.50909 0.50710 0.98317
KNNW 51 0.18293 0.17961 0.07456 0.07080 0.19324 0.18996 0.93851
KNNW 100 0.17480 0.17145 0.12226 0.11870 0.24296 0.23989 0.97406
LOF 15 0.01678 0.01279 0.00765 0.00362 0.03150 0.02757 0.57863
LOF 39 0.00000 -0.00406 0.00786 0.00383 0.03298 0.02906 0.58718
LOF 100 0.00000 -0.00406 0.01041 0.00640 0.03122 0.02729 0.81199
SimplifiedLOF 2 0.02762 0.02368 0.00676 0.00273 0.03289 0.02897 0.53473
SimplifiedLOF 6 0.01303 0.00902 0.00663 0.00260 0.03550 0.03159 0.54504
SimplifiedLOF 34 0.00000 -0.00406 0.00785 0.00383 0.02986 0.02592 0.61273
LoOP 24 0.03659 0.03267 0.00961 0.00559 0.03838 0.03448 0.62219
LoOP 34 0.02846 0.02451 0.01082 0.00680 0.03706 0.03315 0.67287
LoOP 56 0.01220 0.00818 0.01102 0.00700 0.04128 0.03738 0.64492
LoOP 58 0.01220 0.00818 0.01105 0.00704 0.04029 0.03640 0.64621
ODIN 52 0.04065 0.03676 0.01099 0.00698 0.04073 0.03684 0.66816
ODIN 100 0.03080 0.02687 0.01574 0.01175 0.05236 0.04851 0.80159
KDEOS 46 0.04065 0.03676 0.01043 0.00641 0.06082 0.05701 0.56394
KDEOS 50 0.04472 0.04084 0.00973 0.00571 0.04921 0.04535 0.58472
KDEOS 59 0.05691 0.05308 0.01084 0.00682 0.06000 0.05618 0.56994
KDEOS 60 0.05691 0.05308 0.01096 0.00695 0.06048 0.05667 0.57063
LDF 5 0.02385 0.01989 0.00562 0.00159 0.03287 0.02894 0.48184
LDF 97 0.00117 -0.00289 0.02406 0.02009 0.07714 0.07339 0.88444
LDF 100 0.00114 -0.00292 0.02380 0.01984 0.07390 0.07014 0.88705
INFLO 15 0.01678 0.01279 0.00751 0.00348 0.03049 0.02655 0.58353
INFLO 25 0.00000 -0.00406 0.00845 0.00443 0.03560 0.03168 0.63354
INFLO 27 0.00000 -0.00406 0.00847 0.00445 0.03550 0.03159 0.62853
INFLO 100 0.00000 -0.00406 0.00672 0.00269 0.02416 0.02020 0.64210
COF 23 0.02439 0.02043 0.00712 0.00309 0.02592 0.02196 0.54562
COF 35 0.01220 0.00818 0.00876 0.00473 0.03607 0.03216 0.60366
COF 64 0.01220 0.00818 0.00809 0.00406 0.04145 0.03756 0.50673

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, idf weighted categorial attributes

This version contains 40 attributes, 48113 objects, 200 outliers (0.42%)

Download raw algorithm results (328.2 MB) Download raw algorithm evaluation table (58.6 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 0.09500 0.09122 0.01592 0.01181 0.09645 0.09268 0.33270
KNN 100 0.01500 0.01089 0.00992 0.00579 0.03974 0.03573 0.61979
KNNW 1 0.07000 0.06612 0.01692 0.01282 0.09218 0.08839 0.33501
KNNW 2 0.09000 0.08620 0.01632 0.01221 0.09476 0.09098 0.33111
KNNW 3 0.09000 0.08620 0.01577 0.01166 0.09574 0.09197 0.35409
KNNW 100 0.01500 0.01089 0.00930 0.00516 0.05022 0.04626 0.59804
LOF 11 0.10500 0.10126 0.03606 0.03204 0.12830 0.12466 0.57590
LOF 13 0.11000 0.10628 0.03558 0.03156 0.12800 0.12436 0.59344
LOF 14 0.10000 0.09624 0.03543 0.03140 0.13115 0.12752 0.59811
LOF 100 0.07000 0.06612 0.01548 0.01137 0.08387 0.08005 0.64777
SimplifiedLOF 12 0.10500 0.10126 0.03541 0.03138 0.13846 0.13487 0.56678
SimplifiedLOF 15 0.12000 0.11633 0.03687 0.03285 0.12903 0.12540 0.57943
SimplifiedLOF 16 0.12000 0.11633 0.03707 0.03305 0.12955 0.12592 0.58588
SimplifiedLOF 25 0.11500 0.11131 0.03159 0.02755 0.12167 0.11801 0.60154
LoOP 33 0.11500 0.11131 0.03428 0.03025 0.12571 0.12206 0.60653
LoOP 39 0.11500 0.11131 0.03166 0.02762 0.13907 0.13548 0.61047
LoOP 41 0.12000 0.11633 0.03200 0.02796 0.13245 0.12883 0.60491
LoOP 100 0.08000 0.07616 0.02294 0.01887 0.09677 0.09300 0.61454
ODIN 53 0.12000 0.11633 0.02305 0.01897 0.12000 0.11633 0.57948
ODIN 69 0.10792 0.10420 0.02653 0.02247 0.12008 0.11641 0.61668
ODIN 100 0.10812 0.10440 0.02757 0.02351 0.11724 0.11356 0.66154
KDEOS 4 0.03500 0.03097 0.00820 0.00406 0.03911 0.03510 0.49466
KDEOS 36 0.00500 0.00085 0.00592 0.00177 0.02465 0.02058 0.54341
KDEOS 69 0.02000 0.01591 0.01237 0.00824 0.02640 0.02234 0.47988
LDF 10 0.10000 0.09624 0.03550 0.03148 0.12167 0.11801 0.56949
LDF 12 0.10000 0.09624 0.03094 0.02689 0.13333 0.12972 0.57005
LDF 16 0.11500 0.11131 0.02782 0.02376 0.11558 0.11189 0.56735
LDF 100 0.03851 0.03450 0.01783 0.01373 0.05469 0.05074 0.83457
INFLO 13 0.11500 0.11131 0.03838 0.03437 0.12969 0.12606 0.55872
INFLO 24 0.11000 0.10628 0.03939 0.03538 0.13688 0.13328 0.62368
INFLO 32 0.10000 0.09624 0.03430 0.03027 0.12048 0.11681 0.63026
COF 5 0.07000 0.06612 0.02622 0.02216 0.07514 0.07128 0.55921
COF 12 0.09500 0.09122 0.04441 0.04042 0.10256 0.09882 0.54778
COF 18 0.10000 0.09624 0.03926 0.03525 0.10682 0.10310 0.54040
COF 25 0.09500 0.09122 0.03221 0.02817 0.12186 0.11820 0.49243

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, 1-of-n encoding

This version contains 79 attributes, 48113 objects, 200 outliers (0.42%)

Download raw algorithm results (326.9 MB) Download raw algorithm evaluation table (60.4 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 1 0.09500 0.09122 0.01590 0.01179 0.09645 0.09268 0.32708
KNN 100 0.01500 0.01089 0.00990 0.00577 0.03974 0.03573 0.61779
KNNW 1 0.07000 0.06612 0.01690 0.01279 0.09218 0.08839 0.32907
KNNW 2 0.09000 0.08620 0.01629 0.01219 0.09476 0.09098 0.32527
KNNW 3 0.09000 0.08620 0.01575 0.01164 0.09574 0.09197 0.34863
KNNW 100 0.01500 0.01089 0.00928 0.00514 0.05022 0.04626 0.59498
LOF 10 0.10500 0.10126 0.03505 0.03103 0.12879 0.12515 0.56003
LOF 11 0.10500 0.10126 0.03617 0.03214 0.12928 0.12564 0.57123
LOF 14 0.10000 0.09624 0.03550 0.03148 0.13061 0.12698 0.60367
LOF 100 0.07000 0.06612 0.01543 0.01132 0.08387 0.08005 0.64623
SimplifiedLOF 12 0.10500 0.10126 0.03466 0.03063 0.13793 0.13433 0.55961
SimplifiedLOF 15 0.12000 0.11633 0.03659 0.03257 0.12928 0.12564 0.57580
SimplifiedLOF 16 0.12000 0.11633 0.03678 0.03276 0.12979 0.12616 0.58092
SimplifiedLOF 29 0.11000 0.10628 0.03161 0.02757 0.12371 0.12005 0.60012
LoOP 33 0.11500 0.11131 0.03422 0.03019 0.12536 0.12171 0.60910
LoOP 39 0.11500 0.11131 0.03165 0.02761 0.13907 0.13548 0.61018
LoOP 40 0.11500 0.11131 0.03185 0.02780 0.13378 0.13016 0.61185
LoOP 41 0.12000 0.11633 0.03203 0.02798 0.13245 0.12883 0.61134
ODIN 53 0.12000 0.11633 0.02303 0.01895 0.12000 0.11633 0.57582
ODIN 69 0.10778 0.10405 0.02647 0.02240 0.12008 0.11641 0.61385
ODIN 100 0.10661 0.10288 0.02742 0.02336 0.11724 0.11356 0.65799
KDEOS 12 0.03500 0.03097 0.00730 0.00316 0.03825 0.03424 0.51336
KDEOS 36 0.00500 0.00085 0.00583 0.00168 0.02452 0.02045 0.54237
KDEOS 69 0.02000 0.01591 0.01234 0.00821 0.02640 0.02234 0.47891
LDF 9 0.05185 0.04789 0.01400 0.00988 0.06942 0.06554 0.56105
LDF 31 0.05000 0.04603 0.01472 0.01061 0.07436 0.07050 0.60774
LDF 87 0.01653 0.01242 0.01626 0.01216 0.04612 0.04214 0.82400
LDF 99 0.01932 0.01523 0.01727 0.01317 0.05377 0.04982 0.81754
INFLO 13 0.11500 0.11131 0.03874 0.03473 0.12969 0.12606 0.56319
INFLO 24 0.11000 0.10628 0.03952 0.03552 0.13636 0.13276 0.63022
INFLO 25 0.11000 0.10628 0.03959 0.03558 0.13333 0.12972 0.62505
INFLO 32 0.10000 0.09624 0.03433 0.03030 0.12048 0.11681 0.63287
COF 5 0.07000 0.06612 0.02624 0.02217 0.07514 0.07128 0.56094
COF 12 0.09500 0.09122 0.04377 0.03978 0.10224 0.09849 0.54558
COF 18 0.10000 0.09624 0.03841 0.03439 0.10588 0.10215 0.52514
COF 25 0.09500 0.09122 0.03090 0.02686 0.11888 0.11520 0.47691

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, categorial attributes removed

This version contains 40 attributes, 48113 objects, 200 outliers (0.42%)

Download raw algorithm results (328.2 MB) Download raw algorithm evaluation table (58.9 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 0.09500 0.09122 0.01592 0.01181 0.09645 0.09268 0.33270
KNN 100 0.01500 0.01089 0.00992 0.00579 0.03974 0.03573 0.61979
KNNW 1 0.07000 0.06612 0.01692 0.01282 0.09218 0.08839 0.33501
KNNW 2 0.09000 0.08620 0.01632 0.01221 0.09476 0.09098 0.33111
KNNW 3 0.09000 0.08620 0.01577 0.01166 0.09574 0.09197 0.35409
KNNW 100 0.01500 0.01089 0.00930 0.00516 0.05022 0.04626 0.59804
LOF 11 0.10500 0.10126 0.03606 0.03204 0.12830 0.12466 0.57590
LOF 13 0.11000 0.10628 0.03558 0.03156 0.12800 0.12436 0.59344
LOF 14 0.10000 0.09624 0.03543 0.03140 0.13115 0.12752 0.59811
LOF 100 0.07000 0.06612 0.01548 0.01137 0.08387 0.08005 0.64777
SimplifiedLOF 12 0.10500 0.10126 0.03541 0.03138 0.13846 0.13487 0.56678
SimplifiedLOF 15 0.12000 0.11633 0.03687 0.03285 0.12903 0.12540 0.57943
SimplifiedLOF 16 0.12000 0.11633 0.03707 0.03305 0.12955 0.12592 0.58588
SimplifiedLOF 25 0.11500 0.11131 0.03159 0.02755 0.12167 0.11801 0.60154
LoOP 33 0.11500 0.11131 0.03428 0.03025 0.12571 0.12206 0.60653
LoOP 39 0.11500 0.11131 0.03166 0.02762 0.13907 0.13548 0.61047
LoOP 41 0.12000 0.11633 0.03200 0.02796 0.13245 0.12883 0.60491
LoOP 100 0.08000 0.07616 0.02294 0.01887 0.09677 0.09300 0.61454
ODIN 53 0.12000 0.11633 0.02305 0.01897 0.12000 0.11633 0.57948
ODIN 69 0.10792 0.10420 0.02653 0.02247 0.12008 0.11641 0.61668
ODIN 100 0.10812 0.10440 0.02757 0.02351 0.11724 0.11356 0.66154
KDEOS 4 0.03500 0.03097 0.00820 0.00406 0.03911 0.03510 0.49466
KDEOS 36 0.00500 0.00085 0.00592 0.00177 0.02465 0.02058 0.54341
KDEOS 69 0.02000 0.01591 0.01237 0.00824 0.02640 0.02234 0.47988
LDF 10 0.10000 0.09624 0.03550 0.03148 0.12167 0.11801 0.56949
LDF 12 0.10000 0.09624 0.03094 0.02689 0.13333 0.12972 0.57005
LDF 16 0.11500 0.11131 0.02782 0.02376 0.11558 0.11189 0.56735
LDF 100 0.03851 0.03450 0.01783 0.01373 0.05469 0.05074 0.83457
INFLO 13 0.11500 0.11131 0.03838 0.03437 0.12969 0.12606 0.55872
INFLO 24 0.11000 0.10628 0.03939 0.03538 0.13688 0.13328 0.62368
INFLO 32 0.10000 0.09624 0.03430 0.03027 0.12048 0.11681 0.63026
COF 5 0.07000 0.06612 0.02622 0.02216 0.07514 0.07128 0.55921
COF 12 0.09500 0.09122 0.04441 0.04042 0.10256 0.09882 0.54778
COF 18 0.10000 0.09624 0.03926 0.03525 0.10682 0.10310 0.54040
COF 25 0.09500 0.09122 0.03221 0.02817 0.12186 0.11820 0.49243

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, duplicates, idf weighted categorial attributes

This version contains 41 attributes, 60839 objects, 246 outliers (0.40%)

Download raw algorithm results (344.0 MB) Download raw algorithm evaluation table (63.1 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 0.07724 0.07349 0.01376 0.00976 0.08636 0.08265 0.41051
KNN 100 0.01220 0.00818 0.01051 0.00649 0.03758 0.03367 0.71484
KNNW 1 0.07724 0.07349 0.01450 0.01050 0.08440 0.08069 0.39280
KNNW 3 0.07724 0.07349 0.01355 0.00954 0.08531 0.08159 0.41632
KNNW 100 0.01220 0.00818 0.00777 0.00374 0.04683 0.04296 0.57918
LOF 15 0.04956 0.04570 0.01847 0.01448 0.10571 0.10208 0.63858
LOF 16 0.05199 0.04814 0.01914 0.01516 0.10423 0.10059 0.64326
LOF 28 0.00000 -0.00406 0.01209 0.00808 0.05870 0.05487 0.65797
SimplifiedLOF 15 0.04956 0.04570 0.01742 0.01343 0.09987 0.09621 0.59230
SimplifiedLOF 16 0.04816 0.04429 0.01778 0.01379 0.10437 0.10074 0.59092
SimplifiedLOF 33 0.00000 -0.00406 0.01228 0.00827 0.05867 0.05484 0.60617
LoOP 59 0.11382 0.11022 0.02583 0.02188 0.11538 0.11179 0.64334
LoOP 60 0.10976 0.10614 0.02613 0.02218 0.11613 0.11254 0.64537
LoOP 63 0.10976 0.10614 0.02583 0.02188 0.11714 0.11355 0.64683
LoOP 81 0.08537 0.08165 0.02361 0.01965 0.10985 0.10623 0.65046
ODIN 24 0.03961 0.03571 0.01246 0.00845 0.05524 0.05141 0.66386
ODIN 99 0.09593 0.09226 0.02405 0.02009 0.11175 0.10815 0.65440
ODIN 100 0.09793 0.09427 0.02427 0.02031 0.11175 0.10815 0.65441
KDEOS 3 0.04500 0.04112 0.01055 0.00653 0.06428 0.06048 0.50999
KDEOS 4 0.04878 0.04492 0.00862 0.00459 0.05184 0.04799 0.51692
KDEOS 45 0.01220 0.00818 0.00589 0.00185 0.01763 0.01364 0.55541
LDF 24 0.09756 0.09390 0.02152 0.01755 0.10830 0.10468 0.63018
LDF 37 0.09350 0.08982 0.02556 0.02161 0.10147 0.09783 0.65564
LDF 53 0.09889 0.09523 0.02371 0.01974 0.10060 0.09695 0.68753
LDF 67 0.04615 0.04228 0.01973 0.01575 0.07843 0.07469 0.69977
INFLO 15 0.04956 0.04570 0.01839 0.01441 0.10884 0.10523 0.60564
INFLO 16 0.05199 0.04814 0.01901 0.01503 0.10864 0.10502 0.60964
INFLO 34 0.00000 -0.00406 0.01150 0.00749 0.06221 0.05840 0.62712
COF 7 0.05691 0.05308 0.01190 0.00789 0.06588 0.06209 0.55826
COF 13 0.08537 0.08165 0.01755 0.01356 0.09634 0.09267 0.53981
COF 18 0.08130 0.07757 0.01898 0.01499 0.09145 0.08776 0.55255
COF 21 0.08537 0.08165 0.01822 0.01423 0.09810 0.09444 0.52419

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, duplicates, 1-of-n encoding

This version contains 79 attributes, 60839 objects, 246 outliers (0.40%)

Download raw algorithm results (342.4 MB) Download raw algorithm evaluation table (63.7 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 0.07724 0.07349 0.01374 0.00974 0.08636 0.08265 0.40625
KNN 100 0.01220 0.00818 0.01049 0.00647 0.03758 0.03367 0.71323
KNNW 1 0.07724 0.07349 0.01448 0.01048 0.08440 0.08069 0.38847
KNNW 3 0.07724 0.07349 0.01353 0.00952 0.08531 0.08159 0.41185
KNNW 100 0.01220 0.00818 0.00775 0.00372 0.04683 0.04296 0.57668
LOF 15 0.04928 0.04542 0.01850 0.01451 0.10541 0.10178 0.64198
LOF 16 0.05183 0.04798 0.01912 0.01514 0.10408 0.10044 0.64558
LOF 28 0.00000 -0.00406 0.01203 0.00802 0.05719 0.05336 0.66042
SimplifiedLOF 15 0.04928 0.04542 0.01721 0.01322 0.09987 0.09621 0.58856
SimplifiedLOF 16 0.04762 0.04375 0.01750 0.01351 0.10393 0.10029 0.58880
SimplifiedLOF 33 0.00000 -0.00406 0.01218 0.00816 0.05851 0.05469 0.60520
LoOP 57 0.10569 0.10206 0.02411 0.02015 0.10989 0.10628 0.63918
LoOP 61 0.10163 0.09798 0.02526 0.02130 0.11136 0.10775 0.64102
LoOP 62 0.10163 0.09798 0.02525 0.02129 0.11236 0.10876 0.64005
LoOP 81 0.07724 0.07349 0.02289 0.01892 0.10274 0.09910 0.64489
ODIN 18 0.03555 0.03163 0.01051 0.00650 0.03936 0.03546 0.66218
ODIN 60 0.08884 0.08514 0.02141 0.01744 0.11402 0.11042 0.60343
ODIN 97 0.09673 0.09306 0.02431 0.02035 0.11132 0.10772 0.64815
ODIN 100 0.09662 0.09295 0.02452 0.02056 0.11273 0.10913 0.65082
KDEOS 3 0.04104 0.03715 0.00979 0.00577 0.06196 0.05815 0.51023
KDEOS 19 0.02439 0.02043 0.00703 0.00299 0.03657 0.03266 0.55300
LDF 61 0.03898 0.03507 0.01724 0.01325 0.06119 0.05738 0.69296
LDF 75 0.04560 0.04172 0.01959 0.01561 0.07267 0.06890 0.68411
INFLO 16 0.05183 0.04798 0.01900 0.01502 0.10833 0.10471 0.61195
INFLO 34 0.00000 -0.00406 0.01147 0.00745 0.06334 0.05954 0.62899
COF 8 0.06098 0.05716 0.01244 0.00843 0.07179 0.06803 0.56746
COF 18 0.08943 0.08573 0.01868 0.01470 0.09766 0.09399 0.55447
COF 19 0.09756 0.09390 0.01865 0.01466 0.09959 0.09593 0.53336
COF 20 0.09756 0.09390 0.01823 0.01424 0.10191 0.09826 0.51843

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, duplicates, categorial attributes removed

This version contains 38 attributes, 60839 objects, 246 outliers (0.40%)

Download raw algorithm results (344.0 MB) Download raw algorithm evaluation table (63.4 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 1 0.07724 0.07349 0.01374 0.00974 0.08636 0.08265 0.40573
KNN 100 0.01220 0.00818 0.01049 0.00647 0.03758 0.03367 0.71306
KNNW 1 0.07724 0.07349 0.01448 0.01048 0.08440 0.08069 0.38782
KNNW 3 0.07724 0.07349 0.01352 0.00952 0.08531 0.08159 0.41148
KNNW 100 0.01220 0.00818 0.00774 0.00372 0.04683 0.04296 0.57645
LOF 15 0.04928 0.04542 0.01849 0.01451 0.10541 0.10178 0.64233
LOF 16 0.05183 0.04798 0.01914 0.01516 0.10408 0.10044 0.64658
LOF 28 0.00000 -0.00406 0.01197 0.00796 0.05498 0.05115 0.66097
SimplifiedLOF 15 0.04928 0.04542 0.01720 0.01321 0.09987 0.09621 0.58816
SimplifiedLOF 16 0.04762 0.04375 0.01749 0.01351 0.10393 0.10029 0.58873
SimplifiedLOF 33 0.00000 -0.00406 0.01214 0.00813 0.05507 0.05124 0.60514
LoOP 15 0.10163 0.09798 0.02213 0.01816 0.11192 0.10832 0.60074
LoOP 57 0.10569 0.10206 0.02368 0.01972 0.10917 0.10555 0.63918
LoOP 61 0.10163 0.09798 0.02482 0.02086 0.11111 0.10750 0.64095
LoOP 81 0.07724 0.07349 0.02271 0.01875 0.10274 0.09910 0.64493
ODIN 24 0.03974 0.03584 0.01268 0.00867 0.05799 0.05417 0.66182
ODIN 60 0.08897 0.08527 0.02143 0.01745 0.11402 0.11042 0.60341
ODIN 97 0.09673 0.09306 0.02385 0.01989 0.11132 0.10772 0.64801
ODIN 100 0.09662 0.09295 0.02405 0.02009 0.10929 0.10567 0.65063
KDEOS 3 0.04630 0.04243 0.01033 0.00631 0.06601 0.06222 0.51143
KDEOS 19 0.02033 0.01635 0.00701 0.00298 0.03661 0.03270 0.55308
LDF 34 0.08943 0.08573 0.02717 0.02322 0.10582 0.10219 0.65333
LDF 46 0.09756 0.09390 0.02313 0.01916 0.09836 0.09470 0.66789
LDF 72 0.03840 0.03450 0.01940 0.01542 0.07339 0.06963 0.69616
INFLO 16 0.05183 0.04798 0.01903 0.01505 0.10864 0.10502 0.61742
INFLO 34 0.00000 -0.00406 0.01139 0.00738 0.06027 0.05645 0.62897
COF 15 0.08537 0.08165 0.01797 0.01399 0.08915 0.08546 0.56759
COF 18 0.09350 0.08982 0.01865 0.01466 0.09728 0.09361 0.54790
COF 19 0.09756 0.09390 0.01858 0.01460 0.09917 0.09552 0.52656
COF 20 0.09756 0.09390 0.01811 0.01412 0.10169 0.09805 0.51088

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