Sentiment analysis for 18th century multilingual corpora
- Hosting organisations
- Zentrum für Informationsmodellierung - Austrian Centre for Digital Humanities (ZIM-ACDH), Institut für Romanistik, TU Graz - Know-Center, and TU Graz - Institute of Interactive Systems and Data Science
- Responsible persons
- Bernhard Geiger, Elisabeth Hobisch, Philipp Koncar, Martina Scholger, and Denis Helic
- Start
- End
The currently ongoing project Distant Reading for Periodicals of the Enlightenment (DiSpecs - http://gams.uni-graz.at/dispecs ) funded by the Austrian Academy of Sciences analyzes the Spectator periodicals from the Digital Edition project The Spectators in the International Context (Ertler et al. 2011) using different computational methods.
A further expansion of the project, funded by CLARIAH-AT, aims at developing and disseminating a tool chain for sentiment analysis that is applicable to multilingual text corpora of the 18th century. With a focus on lexicon-based sentiment analysis, the first step towards this goal is the annotation of 18th century texts in terms of “polarity”, i.e., the degree to which individual words in the text are perceived as positive or negative. This annotation, provided by experts in the field of Romance studies, is the basis for the creation of new sentiment lexica in the respective language. The thus obtained lexica are subsequently used within a sentiment analysis tool chain that is state-of-the-art in natural language processing tasks, but which has so far rarely been applied to the field of digital humanities.
The tool chain is published in a GitHub repository ( https://github.com/philkon/sentiment-tool-chain ) and comprises two different parts: (i) the creation of sentiment dictionaries and (ii) the actual sentiment analysis. The first part - which can be skipped if dictionaries are already available - provides an approach to create sentiment dictionaries that are applicable to the languages of the 18th century. The second part provides ready-to-use sentiment dictionaries for French, Italian and Spanish as well as methods to analyze sentiment in texts fitting the pertinent corpus.
Publications
Ertler, Klaus-Dieter; Fuchs, Alexandra; Fischer-Pernkopf, Michaela; Hobisch, Elisabeth; Scholger, Martina; Völkl, Yvonne (2011-2021) The Spectators in the international context. https://gams.uni-graz.at/spectators .
Koncar, Philipp; Fuchs, Alexandra; Hobisch, Elisabeth; Geiger, Bernhard; Scholger, Martina; Helić, Denis (2020): Text sentiment in the Age of Enlightenment: an analysis of spectator periodicals. In: Applied Network Science, 5(1), pp. 1-32.
Project team
- Bernhard Geiger
- Christina Glatz
- Denis Helic
- Elisabeth Hobisch
- Philipp Koncar
- Martina Scholger
- Yvonne Völkl
Annotation
- Lena Druml
- Klaus-Dieter Ertler
- Alexandra Fuchs
- Christina Glatz
- Elisabeth Hobisch
- Pia Mayer
- Yvonne Völkl