Comparison of Fuzzy and Traditional Lingusitic Summaries: Use Case Electronic Voting

Supervisors: Narek Andreasyan, Edy Portmann

Contact person: Narek Andreasyan

Student: Looking for student

Project status: Open

Year: 2025

Project Abstract

This master’s thesis introduces a novel approach to evaluating digital ethical issues through fuzzy logic–based linguistic summarization, with electronic voting (e-voting) as the central use case. The research places particular emphasis on a systematic comparison between fuzzy linguistic summaries and traditional summarization methods, demonstrating how the former provide greater flexibility in handling uncertainty, partial truths, and ethical trade-offs. Unlike traditional summaries, which often result in rigid or binary statements, fuzzy linguistic summaries generate nuanced natural language expressions that make complex ethical information more explainable and accessible. A prototype system was developed to operationalize this approach, turning ethically relevant data into user-friendly linguistic statements. As an optional extension, the study also considers large language model (LLM)–based summaries as a complementary point of reference for future work.

To assess the effectiveness of the approach, both qualitative methods and user-centered evaluation were employed. A focus group and structured questionnaire provided insights into usability, interpretability, and perceived usefulness. The evaluation confirmed that participants found fuzzy linguistic summaries more effective than traditional approaches for communicating complex ethical dimensions in e-voting. Moreover, user feedback suggested avenues for refinement in usability and opportunities for expanding the linguistic summarization framework.

Required Skills: Python, JavaScript, Django Framework, HTML, CSS, React.JS

Keywords: Fuzzy Logic; Linguistic summarization; Explainability; Interpretability; Transparency; Digital Services; Ethical Decision-Making; E-Voting; traditional Linuistic summaries.

Document: Not yet available