Classifying and explaining defects with small data for the semiconductor industry
Résumé
In this work, we present an automatic classifier of wafer defects for the semiconductor industry. Hopefully defects are rare, but this puts the classifying problem in a small data context. We propose a fast and fully reproducible approach based on decision trees. The main interest of using decision trees lies in obtaining a highly explicable classifier, which makes the origin of the defect easy to identify.
| Origine | Fichiers produits par l'(les) auteur(s) |
|---|