Abstract
Outlier detection and ensemble learning are well established research directions in data mining yet the application of ensemble techniques to outlier detection has been rarely studied. Building an ensemble requires learning of diverse models and combining these diverse models in an appropriate way. We propose data perturbation as a new technique to induce diversity in individual outlier detectors as well as a rank accumulation method for the combination of the individual outlier rankings in order to construct an outlier detection ensemble. In an extensive evaluation, we study the impact, potential, and shortcomings of this new approach for outlier detection ensembles. We show that this ensemble can significantly improve over weak performing base methods.
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the 26th International Conference on Scientific and Statistical Database Management |
| Number of pages | 12 |
| Publisher | Association for Computing Machinery |
| Publication date | 2014 |
| Article number | 13 |
| ISBN (Print) | 978-1-4503-2722-0 |
| DOIs | |
| Publication status | Published - 2014 |
| Externally published | Yes |
| Event | 26th International Conference on Scientific and Statistical Database Management - Aalborg, Denmark Duration: 30. Jun 2014 → 2. Jul 2014 |
Conference
| Conference | 26th International Conference on Scientific and Statistical Database Management |
|---|---|
| Country | Denmark |
| City | Aalborg |
| Period | 30/06/2014 → 02/07/2014 |
| Sponsor | Danish Otto Monsted Foundation, TARGIT |
Fingerprint
Keywords
- Ensemble
- Outlier detection
Cite this
}
Data perturbation for outlier detection ensembles. / Zimek, Arthur; Campello, Ricardo J.G.B.; Sander, Jörg.
Proceedings of the 26th International Conference on Scientific and Statistical Database Management. Association for Computing Machinery, 2014. 13.Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review
TY - GEN
T1 - Data perturbation for outlier detection ensembles
AU - Zimek, Arthur
AU - Campello, Ricardo J.G.B.
AU - Sander, Jörg
PY - 2014
Y1 - 2014
N2 - Outlier detection and ensemble learning are well established research directions in data mining yet the application of ensemble techniques to outlier detection has been rarely studied. Building an ensemble requires learning of diverse models and combining these diverse models in an appropriate way. We propose data perturbation as a new technique to induce diversity in individual outlier detectors as well as a rank accumulation method for the combination of the individual outlier rankings in order to construct an outlier detection ensemble. In an extensive evaluation, we study the impact, potential, and shortcomings of this new approach for outlier detection ensembles. We show that this ensemble can significantly improve over weak performing base methods.
AB - Outlier detection and ensemble learning are well established research directions in data mining yet the application of ensemble techniques to outlier detection has been rarely studied. Building an ensemble requires learning of diverse models and combining these diverse models in an appropriate way. We propose data perturbation as a new technique to induce diversity in individual outlier detectors as well as a rank accumulation method for the combination of the individual outlier rankings in order to construct an outlier detection ensemble. In an extensive evaluation, we study the impact, potential, and shortcomings of this new approach for outlier detection ensembles. We show that this ensemble can significantly improve over weak performing base methods.
KW - Ensemble
KW - Outlier detection
U2 - 10.1145/2618243.2618257
DO - 10.1145/2618243.2618257
M3 - Article in proceedings
AN - SCOPUS:84904429012
SN - 978-1-4503-2722-0
BT - Proceedings of the 26th International Conference on Scientific and Statistical Database Management
PB - Association for Computing Machinery
ER -