In 2020, the Stanza tool was trained and tested on TalbankenSBX (following MambaDep-style annotation) in order to create a high-quality analysis. Currently (in 2024), this is the default analysis for Swedish in Sparv
This analysis is used with Sparv. Check out Sparv's quick start guide to get started!
To use this analysis, add the following lines under export.annotations in the Sparv corpus configuration file:
- <token>:stanza.dephead_ref # Sentence-relative positions of the dependency heads
- <token>:stanza.deprel # Dependency relations to the head
- <token>:stanza.ref # Token IDs relative to their sentences
For more info on how to use Sparv, check out the Sparv documentation.
Example output:
<token dephead_ref="3" deprel="SS" ref="1">Det</token>
<token dephead_ref="1" deprel="HD" ref="2">här</token>
<token deprel="ROOT" ref="3">är</token>
<token dephead_ref="5" deprel="DT" ref="4">en</token>
<token dephead_ref="3" deprel="SP" ref="5">korpus</token>
<token dephead_ref="3" deprel="IP" ref="6">.</token>
A model trained on TalbankenSBX_train and validated on TalbankenSBX_dev yields Labelled Attachment Score of 84.48 on TalbankenSBX_test.
Stanza: Peng Qi, Yuhao Zhang, Yuhui Zhang, Jason Bolton and Christopher D. Manning. 2020
Stanza: A Python Natural Language Processing Toolkit for Many Human Languages. In Association for Computational Linguistics (ACL) System Demonstrations. 2020
SUC3: https://spraakbanken.gu.se/en/resources/suc3
TalbankenSBX: https://spraakbanken.gu.se/en/blog/20200609-the-five-lives-of-talbanken
SIC2: https://spraakbanken.gu.se/en/resources/sic2