List your company
ResearchFormerly WizWand

Anomaly Detection on ADGym Average performance across 14 domains

75.9AUCPRGround Truth (Best Design)
34.92445.56256.266.838Sep 27, 2023
Updated 8mo ago

Evaluation results

MethodLinks
Ground Truth (Best Design)Number of labeled anomalies=202023.0975.9
Ground Truth (Best Design)Number of labeled anomalies=102023.0973.7
Ground Truth (Best Design)Number of labeled anomalies=52023.0970.6
ADGym (ML-ensemble)Number of labeled anomalies=20, ML Algorithm=CatBoost2023.0966.9
ADGym (ML-ensemble)Number of labeled anomalies=20, ML Algorithm=XGBoost2023.0966.2
ADGym (ML-single)Number of labeled anomalies=20, ML Algorithm=CatBoost2023.0965.7
ADGym (ML-single)Number of labeled anomalies=20, ML Algorithm=XGBoost2023.0964.2
ADGym (DL-ensemble)Number of labeled anomalies=20, Meta-feature extraction=2-stage2023.0963.9
ADGym (DL-ensemble)Number of labeled anomalies=20, Meta-feature extraction=end2end2023.0962.3
ADGym (DL-single)Number of labeled anomalies=20, Meta-feature extraction=2-stage2023.0962.2
ADGym (ML-ensemble)Number of labeled anomalies=10, ML Algorithm=XGBoost2023.0961.7
ADGym (DL-single)Number of labeled anomalies=20, Meta-feature extraction=end2end2023.0961.6
ADGym (ML-ensemble)Number of labeled anomalies=10, ML Algorithm=CatBoost2023.0961.4
ADGym (ML-single)Number of labeled anomalies=10, ML Algorithm=CatBoost2023.0960.3
ADGym (DL-ensemble)Number of labeled anomalies=10, Meta-feature extraction=2-stage2023.0960
ADGym (ML-single)Number of labeled anomalies=10, ML Algorithm=XGBoost2023.0960
ADGym (DL-ensemble)Number of labeled anomalies=10, Meta-feature extraction=end2end2023.0959.7
ADGym (DL-single)Number of labeled anomalies=10, Meta-feature extraction=end2end2023.0959.2
ADGym (DL-single)Number of labeled anomalies=10, Meta-feature extraction=2-stage2023.0958.9
ADGym (ML-ensemble)Number of labeled anomalies=5, ML Algorithm=XGBoost2023.0955.3
Supervised SelectionNumber of labeled anomalies=202023.0955
ADGym (ML-ensemble)Number of labeled anomalies=5, ML Algorithm=CatBoost2023.0954.7
ADGym (ML-single)Number of labeled anomalies=5, ML Algorithm=CatBoost2023.0953.9
ADGym (DL-ensemble)Number of labeled anomalies=5, Meta-feature extraction=2-stage2023.0953.8
ADGym (DL-ensemble)Number of labeled anomalies=5, Meta-feature extraction=end2end2023.0953.4
ADGym (DL-single)Number of labeled anomalies=5, Meta-feature extraction=end2end2023.0953.3
ADGym (ML-single)Number of labeled anomalies=5, ML Algorithm=XGBoost2023.0953.2
ADGym (DL-single)Number of labeled anomalies=5, Meta-feature extraction=2-stage2023.0953.1
Random SelectionNumber of labeled anomalies=202023.0950.9
Random SelectionNumber of labeled anomalies=102023.0947.8
Supervised SelectionNumber of labeled anomalies=102023.0945.9
Random SelectionNumber of labeled anomalies=52023.0944.7
Supervised SelectionNumber of labeled anomalies=52023.0936.5