| Autor: |
Jelenčič, Jakob, Rožanec, Jože M., Mladenić, Dunja |
| Předmět: |
|
| Zdroj: |
Central European Conference on Information & Intelligent Systems; 2022, p49-54, 6p |
| Abstrakt: |
The main contribution of the paper is method that combines adaptive windowing and the drift detection method for detecting concept drift and therefore determine when a machine learning model should be retrained. We evaluated the method on four real-world datasets with concept drift. Our results show that the proposed method improves the models' performance (in three of four datasets) on unseen data when compared to the baseline. [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
