UD_Slovenian-SSJ

Treebank card from UD release 2.18: metadata and texts come from UD_Slovenian-SSJ/README.md; size is counted by streaming the .conllu files and checked against the release stats.xml. Generated by tools/treebanks build; do not edit by hand. Papers: Papers: UD_Slovenian-SSJ; seeds: seeds/; converter: convert.md.

Annotation layers

UD scale: manual native — annotated by hand directly in UD; converted from manual — converted from manual annotation in another scheme; converted with corrections — converted and partly corrected by hand; automatic with corrections — automatic, with partial manual corrections; automatic — automatic; not available — the layer is absent.

Size

Tokens are surface tokens (Spanish vámonos is one), words are syntactic words (vamos + nos); multiword tokens are those split into several words; empty nodes are the N.M nodes of the enhanced graph.

Summary

The SSJ treebank is the reference UD treebank for Slovenian, consisting of approximately 13,000 sentences and 267,097 tokens from fiction, non-fiction, periodical and Wikipedia texts in standard modern Slovenian. As of UD release 2.10 in May 2022, the original version of the SSJ UD treebank has been partially manually revised and extended with new manually annotated data.

Introduction

The original Slovenian SSJ UD Treebank, first released as part of UD v1.2 in 2015 (Dobrovoljc et al. 2017), was created through a fine-grained rule-based conversion of the ssj500k treebank (Krek et al. 2021), the largest collection of manually syntactically annotated data in Slovenian, originally annotated in the JOS annotation scheme (Erjavec et al. 2010). With the exception of the manual disambiguation of the AUX and VERB occurrences of the verb 'biti' (to be), the conversion was fully automatic, based on a set of rules for both morphosyntactic and syntactic layer, which include different lexical, morphological and dependency features, while the original ssj500k tokenization and lemmatization principles remained unchanged. The rules and conversion scripts are available at github.com/clarinsi/jos2ud.

In 2022, the original SSJ UD treebank was partially manually revised to correct the previously identified annotation inconsistencies, and implement the newly introduced changes in the annotation guidelines. In addition, the treebank was substantially extended to almost double the original size, with new manually annotated sentences coming from the previously unreleased subset of the ssj500k corpus, and the Slovenian subset of the ELEXIS parallel sense-annotated corpus of Wikipedia texts (Martelli et al. 2021). Despite the extension, the data split remained unchanged with the original SSJ sentences being preserved as part of the same train-dev-test subset. More details on the latest SSJ UD version are given in Dobrovoljc and Ljubešić (2022).

The metadata for most documents in the treebank includes also the document genre. The documents fall into six different genre categories: newspaper, magazine, internet, professional, fiction, other.

The issues section of this repository serves as a platform for general discussion regarding suggestions for the Slovenian UD guidelines and other open issues.

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/treebanks/sl/ssj/