Journal for Language Technology and Computational Linguistics (JLCL)

Krause, Thomas

Leser, Ulf

Lüdeling, Anke

English

graphANNIS: A Fast Query Engine for Deeply Annotated Linguistic Corpora

Diewald, Nils

Margaretha, Eliza

Krill: KorAP search and analysis engine

Neumann, Arne

Merging and validating heterogenous, multi-layered corpora with discoursegraphs

Geyken, Alexander

Kupietz, Marc

Editorial
Alexander Geyken, Marc Kupietz
Editorial
With the growing availability and importance of (large) corpora in all fields of linguistics,
the role of software tools is gradually moving from useful, possibly intelligent information-
technological “helpers” towards scientific instruments that are as integral parts of the
research process as data, methodology and interpretations. Both aspects are present
in this special issue of JLCL on corpus linguistic software tools: The contributions
address topics such as software tools for managing corpora, transforming corpus data,
annotating and analyzing corpora as well as innovative ways of exploring and visualizing
corpus data and analyses.
Thomas Krause, Ulf Leser and Anke Lüdeling present graphANNIS, a newly developed
graph database for querying deeply annotated linguistic corpora. After discussing the
pros and cons of existing solutions, they document implementation details, the data
model, and query aspects, including the evaluation of the the proposed graph based
implementation vs. the established RDBMS backend. It is reported that graphANNIS
performs in the majority of the tested cases much faster than the relational equivalent.
Paul Rayson, John Mariani, Bryce Anderson-Cooper, Alistair Baron, David Gullick,
Andrew Moore and Stephen Wattam address in their paper “towards interactive mul-
tidimensional visualisations for corpus linguistics” several important issues, including
how iterative and exploratory corpus investigations can be supported by dynamic and
interactive visualisation techniques and how to approach issues with current corpus
retrieval tools by including interactive and multidimensional visualizations.
Christian Pölitz presents in his paper “data mining software for corpus linguistics with
an application in diachronic linguistics” a freely available plugin for the RapidMiner
software. This plug-in could be of particular relevance for researchers already using
RapidMiner, or for computational linguists looking for a tool that is more user-friendly
- through drag-and-drop - than e.g. the established R software. The author uses the
example case of topic modelling over time to demonstrate the advantages of the new
plugin, including the implemented topic models and coherence measures.
Nils Diewald and Eliza Margaretha present “Krill”, the search component of the KorAP
corpus search and analysis platform. This paper makes three important contributions to
research and software engineering in the area of corpus indexing and query: It introduces
a new open-source corpus indexing software based on Apache Lucene and describes
how linguistic corpus search can be implemented on top of a full text search engine
such as Lucene. Furthermore, as an implementation of the KoralQuery specification,
Krill is an important milestone for the types of search pattern expected from a modern
linguistic corpus query engine.
Maria Skeppstedt, Carita Paradis and Andreas Kerren present PAL, a standalone tool
for pre-annotation and active learning. PAL can be trained using gold standard data
JLCL 2016 – Band 31 (1) iii
Geyken, Kupietz
and it can incorporate manually annotated/corrected data created by the user within
the BRAT annotation tool. PAL tries to optimize the process by using an active
learning approach to select sentences to be annotated by the user. The tool focuses on
"chunk"-based annotation tasks, e.g. for named entity detection.
Arne Neumann presents “discourse graphs”, a library and command-line application for
merging and validating heterogeneous, multi-layered corpora. The resulting software
tool consists of a python library and command-line applications for converting multi-
level annotated data into different formats, merging independent annotations of the
same text into one graph representation and validating heterogeneous annotation layers
of the same text. This complex model is built upon the property graph model and can
therefore benefit from the ecosystem built around modern graph libraries based on that
paradigm.
Thomas Schmidt presents software tools and workflows for the construction and dissem-
ination the FOLK corpus, a corpus of spoken interaction. The article covers the tools
used in the individual steps of transcription, anonymization, orthographic normalization,
lemmatization and POS tagging of the data, as well as the utilities used for corpus
management. Furthermore, it presents the DGD (Datenbank für Gesprochenes Deutsch
- Database of Spoken German) as means for distributing the completed research data
and making it available for qualitative and quantitative analysis.
Ruprecht von Waldenfels and Michał Woźniak present SpoCo, an adaptable query
and analysis system for spoken dialect corpora with aligned audio data encoded in
ELAN. SpoCo is targeted at users of different levels of expertise and provides them
with advanced concordancing functions, as well as the the possibility to edit and correct
transcriptions. SpoCo takes a middle position between systems that are developed for
the purposes of a specific dialect corpus, on the one hand, and general-use systems
designed for a wide range of data and usage cases, on the other. It is used in a network
of Slavic dialect projects that cooperate in tool development and data sharing.
Our gratitude goes to the colleagues who contributed to this special issue as external
reviewers.
Berlin and Mannheim, May 2017
Alexander Geyken
Marc Kupietz
iv JLCL


Editorial

Mittmann, Roland

Automatisierter Abgleich des Lautstandes althochdeutscher Wortformen

von Waldenfels, Ruprecht

Woźniak, Michał

SpoCo – a simple and adaptable web interface for dialect corpora

Skeppstedt, Maria

Paradis, Carita

Kerren, Andreas

PAL, a tool for Pre-annotation and Active Learning

Schmidt, Thomas

Construction and Dissemination of a Corpus of Spoken Interaction – Tools and Workflows in the FOLK project

Petran, Florian

Bollmann, Marcel

Dipper, Stefanie

Klein, Thomas

ReM: A reference corpus of Middle High German -- corpus compilation, annotation, and access

Hoenen, Armin

Samushia, Lela

Gepi: An Epigraphic Corpus for Old Georgian and a Tool Sketch for Aiding Reconstruction

Not a member yet