Hybrid-Based Detection Of Plagiarized Amharic News Articles On Social Media Platforms
Files
Date
2026
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The rapid growth of social media news dissemination in Ethiopia has raised increasing
concerns about plagiarism in Amharic news articles. Such practices undermine journalistic
integrity, reduce originality in reporting, and contribute to public mistrust in news
organizations and media credibility. To address this issue, this study recommends a hybrid
plagiarism detection framework specifically designed for Amharic news content shared on
social media platforms. The research employed purposive sampling to collect data from a
diverse range of sources, including government and private news agencies, individual
bloggers, journalists, and international media outlets such as BBC Amharic. The selected
sources were required to publish daily news content across multiple domains, including
politics, economics, and international affairs, and to maintain active Amharic-language
public channels with significant audience engagement. The proposed framework integrates
both lexical and semantic analysis techniques. The first layer applies traditional candidate
detection methods, including fingerprinting and n-gram analysis, to identify potentially
plagiarized segments. The second layer refines these results using advanced semantic
approaches, specifically Latent Dirichlet Allocation
