Fast or Accurate? – A Comparative Evaluation of PoS Tagging Models
We perform a comparison of 22 PoS tagger models for English and German offered by 9 different implementations. By evaluating on a mix of corpora from different domains, we simulate a black-box usage where researchers select a tagger (because of popularity, ease of use, etc.) and apply it to all sorts of text. We find the expected trade-off between fast models with relatively low accuracy and slower models with higher accuracy. The choice of the model, even for the same tagger, does matter and the model should always be chosen for the task at hand. Our evaluation provides researchers with a basis for selecting taggers according to their needs.
Vorschau
Zitieren
Rechte
Nutzung und Vervielfältigung:
Dieses Werk kann unter einerCreative Commons Namensnennung - Nicht-kommerziell - Weitergabe unter gleichen Bedingungen 4.0 Lizenz (CC BY-NC-SA 4.0)
genutzt werden.