An Automated Analysis Using the ReaderBench Framework Cover Image

An Automated Analysis Using the ReaderBench Framework
An Automated Analysis Using the ReaderBench Framework

Author(s): Simina-Maria Terian, Teodor-Mihai COTEȚ, Maria-Dorinela Sîrbu, Mihai Dascălu, Ştefan Trăuşan-Matu
Subject(s): Language and Literature Studies, Economy, Electronic information storage and retrieval
Published by: Academia Română – Centrul de Studii Transilvane
Keywords: discourse of economic crisis; types of discourse; Romanian media; Natural Language Processing; ReaderBenchFramework; textual complexity indices;

Summary/Abstract: many people are specialists in economics, but everyone tends to have the impression of knowing something about economic crises—and, above all, they are eager to talk about it. The economic crisis seems to have been the dominant topic of the last decade at least in Romania, generating a plurality of parallel or even contradictory discourses. This paper explores, through an automated analysis using the ReaderBenchFramework, Romanian media discourses on the economic crisis spanning the years between 2008 and 2018. We start by delineating four categories of discourse relevant for the selected topic: specialized academic publications (SA), specialized non-academic publications (SN), non-specialized non-academic publications (NN), and informal debate platforms (ID). For each of these categories, we selected 50 texts, which were then analyzed using the ReaderBenchFramework. The 13 statistically significant textual complexity indices that resulted from this automated analysis show that the first three categories of discourse follow a downward progression pattern, while the fourth category (ID) appears to exhibit an eccentric behavior. The central thesis of our article is that this behavior stems from the ID’s attempt to replace the specialized non-acade

  • Issue Year: XXVIII/2019
  • Issue No: Suppl. 1
  • Page Range: 245-262
  • Page Count: 16
  • Language: English