Cerè, Sofia
(2024)
Promoting web accessibility and inclusion of people with intellectual disabilites: natural language processing driven easy-to-read (E2R) transformation for italian written text.
[Laurea magistrale], Università di Bologna, Corso di Studio in
Informatica [LM-DM270], Documento full-text non disponibile
Il full-text non è disponibile per scelta dell'autore.
(
Contatta l'autore)
Abstract
This research addresses the challenge of enhancing accessibility and comprehension of bureaucratic Italian text on Public Administration (PA) websites for individuals with intellectual disabilities (PwID). It focuses on the limited availability of Easy-to-Read (E2R) content, which hampers civil participation. To tackle this issue, the study proposes an automated Natural Language Processing (NLP) tool for lexical simplification, which identifies complex words and replaces them with simpler alternatives. Easy-to-Read language improves clarity through straightforward vocabulary, simple sentence structures, and logically organized content tailored to audience needs. This model benefits PwID and others who find standard language challenging. The research aligns with the United Nations Convention on the Rights of Persons with Disabilities, emphasizing the right to clear information. The study explores three research questions. The first investigates the requirements for a tool supporting E2R transformation through literature reviews and expert interviews, highlighting the need for linguistic simplification while maintaining coherence. The second question assesses the effectiveness of NLP tools in E2R text transformation using a prototype focused on simplifying complex words and enhancing readability. The third question examines the usability of the NLP-generated E2R content through quantitative readability indices and qualitative feedback from PwID. Findings emphasize the necessity of NLP capabilities and evaluation mechanisms for comprehension. Although the prototype showed promise in identifying complex terms, challenges regarding context and vocabulary limitations were noted. Comparisons with advanced language models like ChatGPT revealed that the prototype sometimes fell short in user comprehension, with feedback stressing the importance of familiar terms and preserving meaning in simplifications.
Abstract
This research addresses the challenge of enhancing accessibility and comprehension of bureaucratic Italian text on Public Administration (PA) websites for individuals with intellectual disabilities (PwID). It focuses on the limited availability of Easy-to-Read (E2R) content, which hampers civil participation. To tackle this issue, the study proposes an automated Natural Language Processing (NLP) tool for lexical simplification, which identifies complex words and replaces them with simpler alternatives. Easy-to-Read language improves clarity through straightforward vocabulary, simple sentence structures, and logically organized content tailored to audience needs. This model benefits PwID and others who find standard language challenging. The research aligns with the United Nations Convention on the Rights of Persons with Disabilities, emphasizing the right to clear information. The study explores three research questions. The first investigates the requirements for a tool supporting E2R transformation through literature reviews and expert interviews, highlighting the need for linguistic simplification while maintaining coherence. The second question assesses the effectiveness of NLP tools in E2R text transformation using a prototype focused on simplifying complex words and enhancing readability. The third question examines the usability of the NLP-generated E2R content through quantitative readability indices and qualitative feedback from PwID. Findings emphasize the necessity of NLP capabilities and evaluation mechanisms for comprehension. Although the prototype showed promise in identifying complex terms, challenges regarding context and vocabulary limitations were noted. Comparisons with advanced language models like ChatGPT revealed that the prototype sometimes fell short in user comprehension, with feedback stressing the importance of familiar terms and preserving meaning in simplifications.
Tipologia del documento
Tesi di laurea
(Laurea magistrale)
Autore della tesi
Cerè, Sofia
Relatore della tesi
Scuola
Corso di studio
Indirizzo
CURRICULUM A: TECNICHE DEL SOFTWARE
Ordinamento Cds
DM270
Parole chiave
Easy-to-Read, Intellectual Disabilities, Natural Language Processing, Web Accessibility, Inclusion, Italian Text
Data di discussione della Tesi
30 Ottobre 2024
URI
Altri metadati
Tipologia del documento
Tesi di laurea
(NON SPECIFICATO)
Autore della tesi
Cerè, Sofia
Relatore della tesi
Scuola
Corso di studio
Indirizzo
CURRICULUM A: TECNICHE DEL SOFTWARE
Ordinamento Cds
DM270
Parole chiave
Easy-to-Read, Intellectual Disabilities, Natural Language Processing, Web Accessibility, Inclusion, Italian Text
Data di discussione della Tesi
30 Ottobre 2024
URI
Gestione del documento: