Guest lecture "Learner language analysis with big data: A workshop report from the DAKODA project." | Lisa Lenort

Speaker: Lisa Lenort (Leipzig University)
Location: TU Dresden, Wiener Straße 48, Room W48/0101
Date/Time: 17.01.2024, 14:50 - 16:20

The DAKODA project is dedicated to the computational linguistic analysis of syntactic acquisition levels of German using language data from learners of German as a foreign and second language. For this purpose, many learner corpora of German are collected, consolidated into a large database and prepared for (partially) automatic analyses.
In addition to already available collections such as the corpora of the Falko family or MERLIN, important corpora from classic second language acquisition studies, such as the ESF or ZISA data, as well as previously unpublished corpora, such as MULTILIT, will also be included. Bringing together this sometimes quite inhomogeneous data for cross-corpus analyses poses challenges on several levels. Many of these relate to the processing of the collected learner data itself. In addition to the technical pre-processing of the language data (in terms of formats, annotations, etc.), metadata (about the learners, the setting, the task, etc.) are central to the focused questions. For this reason, DAKODA is developing a metadata schema that links the various corpora together and thus enriches the language data with important information about the speakers (etc.). The presentation will introduce the project, its goals and prerequisites and then focus on the metadata work area. In addition to the first draft of the metadata schema for DAKODA and its genesis, the relevance of metadata for learner corpus research and the associated need for standardization efforts will be emphasized.

DAKODA (Datenkompetenzen in DaF/DaZ: Exploration sprachtechnologischer Ansätze zur Analyse von L2-Erwerbsstufen in Lernerkorpora des Deutschen) is an interdisciplinary project of the Language Technology Lab of the FernUniversität Hagen (Head: Prof. Dr.-Ing. Torsten Zesch) and the Herder Institute of the University of Leipzig (Head: Prof. Dr. Katrin Wisniewski). For more information, see www.dakoda.org.

Interested students and colleagues are cordially invited to attend.

Participation in the guest lecture is free of charge.

Contact: Jupp Möhring (jupp.moehring@tu-dresden.de)

Translated with DeepL.com (free version)