2026-05-09T15:58:49.934Z http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai

ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai:1:10214 2024-12-03T20:10:47Z 1

Natural Language Processing Approach for Fake News Detection Using Metaheuristics Optimized Extreme Gradient Boosting 2024 http://ezaposleni.singidunum.ac.rs/rest/sciNaucniRezultati/oai/record/1/10214 https://ieeexplore.ieee.org/abstract/document/10731062 A. Petrovic J. Perisic L. Jovanovic M. Zivkovic M. Antonijevic N. Bacanin The problem of fake news can be dangerous, as widely spread misinformation can cause serious physical and psychological harm. In the age of digitization and artificial intelligence (AI), particularly deep learning methods, which are able to generated synthetic text, being able to distinguish between real and fake news has become even greater challenge. Therefore, this research proposes a metaheuristics optimized machine learning (ML) approach with natural language processing (NLP) application for detecting fake news. The main classification model that was used is the eXtreme gradient boosting (XGBoost), which was tuned by the modified variable neighborhood search (VNS) algorithm. The NLP technique that was applied to transform text into the meaningful form for ML is the term frequency-inverse document frequency (TF-IDF). The performance of the VNS-tuned XGBoost was evaluated and compared with other cutting edge metaheuristics, where the proposed method obtained superior performance. conferenceObject 252 257 10.1109/AIC61668.2024.10731062 2024 IEEE 3rd World Conference on Applied Intelligence and Computing (AIC), IEEE, Gwalior, India