Abstract
In the field of supervised machine learning, classification is a central task that assigns labels based on their probability distribution. Classification often follows the MAP criterion, but this can be inadequate when categories are ordered, as in satisfaction scales. This article proposes Ord-MAP, an optimal alternative that opens the door to better practices in ordinal classification.
| Translated title of the contribution | What label should we assign? :: Classification under uncertainty |
|---|---|
| Original language | Catalan |
| Pages (from-to) | 0001-2 |
| Number of pages | 2 |
| Journal | Divulga UAB – Revista de Difusió de la Recerca de la Universitat Autònoma de Barcelona |
| Publication status | Published - 2025 |
Fingerprint
Dive into the research topics of 'What label should we assign? : Classification under uncertainty'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver