Abstract | ||
---|---|---|
This work is focused on the study of hybrid one-class classification techniques, used for anomaly detection on a control level plant. The initial dataset is obtained from the system, working at different operating points, corresponding to three opening degrees of the tank drain valve. The issue of working in different plant configurations is solved through a hybrid classifier, achieved using clustering algorithms combined with a one-class boundary method. The hybrid classifier performance is trained, tested and validated by creating real anomalies changing the drain valve operation. The final classifier is validated, with an AUC value 90.210%, which represents a successful performance. |
Year | DOI | Venue |
---|---|---|
2020 | 10.1007/978-3-030-57805-3_27 | CISIS |
DocType | Citations | PageRank |
Conference | 0 | 0.34 |
References | Authors | |
0 | 7 |
Name | Order | Citations | PageRank |
---|---|---|---|
Iago Núñez | 1 | 0 | 0.34 |
Esteban Jove | 2 | 20 | 15.33 |
José Luís Casteleiro-Roca | 3 | 98 | 22.27 |
Héctor Quintián | 4 | 73 | 9.42 |
Francisco Zayas-Gato | 5 | 0 | 0.34 |
Dragan Simic | 6 | 40 | 12.78 |
José Luís Calvo-Rolle | 7 | 175 | 41.67 |