A novel approach based on rough set theory for analyzing information disorder

Area of Science:

• Computational Social Science
• Information Science
• Data Mining

Background:

• Information Disorder, encompassing misinformation and disinformation, poses significant challenges in social media environments.
• Existing analytical methods struggle to fully capture the complex dynamics of information spread and user interactions.
• Rough Set Theory offers a novel framework for analyzing imprecise and uncertain data, relevant to social media phenomena.

Purpose of the Study:

• To present and evaluate an approach using Rough Set Theory (RST) and its extensions for analyzing Information Disorder.
• To model social media users and information circulation using RST concepts like indiscernibility and approximations.
• To contextualize and evaluate system theory concepts (Complexity, Milestone) for Information Disorder using information-theoretic measures.

Main Methods:

• Application of Rough Set Theory concepts: lower/upper approximations, indiscernibility relations, neighborhood relations.
• Utilizing information-theoretic measures: roughness and entropy.
• Developing and applying four algorithms based on RST and its variants/extensions.
• Case study analysis using real-world data related to COVID-19 disinformation.

Main Results:

• Demonstrated the utility of RST constructs to model key elements of Information Disorder (message, interpreters).
• Quantified the interpretative ability of Information Disorder effects using measures like the ratio of lower/upper approximations.
• Showcased improved understanding of data fragmentation and cohesion in information spread.
• Identified the superiority of Fuzzy Rough Sets approaches for interpreting Information Disorder phenomena.

Conclusions:

• Rough Set Theory provides a powerful, novel framework for analyzing Information Disorder in social media.
• The proposed RST-based measures enhance the interpretability of misinformation spread dynamics.
• Fuzzy Rough Sets offer superior performance in understanding complex information disorder patterns, benefiting social media analysts and platform providers.