Predicting Slovene Text Complexity Using Readability Measures Cover Image

Predicting Slovene Text Complexity Using Readability Measures
Predicting Slovene Text Complexity Using Readability Measures

Author(s): Tadej Škvorc, Simon Krek, Senja Pollak, Špela Arhar Holdt, Marko Robnik-Šikonja
Subject(s): Computational linguistics, Slovenian Literature, ICT Information and Communications Technologies, British Literature
Published by: Inštitut za novejšo zgodovino
Keywords: readability; natural language processing; text analysis;

Summary/Abstract: The majority of existing readability measures are designed for English texts. We aim to adapt and test the readability measures on Slovene. We test ten well-known readability formulas and eight additional readability criteria on five types of texts: children’s magazines, general magazines, daily newspapers, technical magazines, and transcriptions of national assembly sessions. As these groups of texts target different audiences, we assume that the differences in writing styles should be reflected in their readability scores. Our analysis shows which readability measures perform well on this task and which fail to distinguish between the groups.

  • Issue Year: 59/2019
  • Issue No: 1
  • Page Range: 198-220
  • Page Count: 23
  • Language: English