Brown Bag Lunch: Simone Rebora (Verona/GCDH) and Svenja Simone Guhr (GCDH): titles see below...

, 12:15 pm to

GCDH, Heyne Haus, Papendiek 16, 37073 Göttingen, Seminarraum 1. 1) S. Rebora, J.B. Herrmann, G. Lauer, M. Salgaro: Using stylometry to grasp the style of "the man without qualities". Robert Musil's activity as editor of the Tiroler Soldaten-Zeitung (1916-1917). 2) Svenja Simone Guhr (GCDH):
Sentiment Analysis of User Comments below Online Newspaper Articles concerning the French Presidential Election in 2017.

_________________________________________________________________________________________

S. Rebora, J.B. Herrmann, G. Lauer, M. Salgaro:

Using stylometry to grasp the style of "the man without qualities". Robert Musil's activity as editor of the Tiroler Soldaten-Zeitung (1916-1917).

_________________________________________________________________________________________

 

During WWI, between 1916 and 1917, Robert Musil was chief editor of the Tiroler Soldaten-Zeitung (TSZ) in Bozen. This activity has posed a philological problem to Musil scholars, who have not been able to attribute with certainty a range of texts to the author. With this paper, we present a new approach that combines historical and philological research with stylometric methods.

First, we developed a combinatory design for tackling the issue of text length: longer chunks composed by the juxtaposition of single texts were analyzed by a selection of 16 classifiers. Second, we explored WWI archives and we digitized historical documents in search for possible candidates for authorship. Results of our experiment suggest that Musil attribution may be disproved with a high level of confidence for 9 texts, that were more probably written by a less-known author, Albert Ritter.

 

_________________________________________________________________________________________

Svenja Simone Guhr (GCDH):

Sentiment Analysis of User Comments below Online Newspaper Articles concerning the French Presidential Election in 2017

_________________________________________________________________________________________

 

Abstract

Social media has been of growing importance in recent elections worldwide. Publicly available online publications from popular newspapers, in particular, have a considerable influence on what the public is informed about. Moreover, they provide an open space for political discussions with regard to electoral issues. Presupposing this, an automatic analysis of comments published below such newspaper articles allows to understand the public sentiment and polarisation with regard to electoral candidates and issues. It might as well serve as a basis for electoral victory prediction for a chosen candidate.

 

Subject to my presentation is my current research project, which involves an automatic annotation of online articles of the French newspaper Le Monde and related user comments carried out in the context of the 2017 French presidential election. My focus lies on an analysis of the related user comments with regard to the expressed sentiments.

In order to present the variety of topics addressed in the online articles concerning the election, I will demonstrate how I singled out a selection of articles by clustering them with the software package stylo for the software environment R. Additionally, I will describe how I prepared the selected user comments for an automatic sentiment analysis.

The study focusses on the use of sentiment indicating adjectives in the user comments. Many French adjectives can be polarised into the two categories positive and negative (neutral for the non-classifiable adjectives). Hence, the question I am going to look at is whether the use of a specific adjective in a user comment indicates the commentator’s consent or disagreement with regard to the topic of the newspaper article.

 

Furthermore, I am going to present my self-developed French sentiment lexicon and a sentiment-tagger created on the base of the lexicon. By applying the sentiment-tagger to the user comments I automatically generated annotated data for the user comments detecting and polarising the French adjectives used therein (positive, negative, neutral).

Eventually, an analysis of the polarisation of the French adjectives used in the user comments will allow to detect the general mood (consent/disagreement) of the commentators with regard to the topics addressed in the related online articles.