EZE, IFEANYI VALENTINE (2026) HYBRID APPROACH FOR INTELLIGENT FAKE NEWS DETECTION USING NATURAL LANGUAGE PROCESSING. Other thesis, Godfrey Okoye University, Enugu Nigeria.
![[img]](https://eprints.gouni.edu.ng/style/images/fileicons/text.png) |
Text
EZE IFEANYI VALENTINE FNAL PROJECT.pdf Download (1MB) |
Abstract
The fast emergence of news websites and social media platforms has contributed to the growth of the fake news problem, making it hard to verify news, build public trust, and make reasonable decisions. Since conventional fact-checking techniques cannot keep up with the pace and quantity of the misinformation dissemination process, there is a need to come up with automated mechanisms for detecting fake news. This paper proposes the implementation of a Hybrid Fake News Detection System that will utilize Natural Language Processing (NLP) and deep learning technologies for classifying news articles as either real or fake. The data used in the experiment included a labeled dataset of both real and fake news articles that was prepared using text cleaning and tokenization. This model combines the capabilities of BERT, which extracts contextual features, and LSTM that learns sequential patterns. The experimental evaluation of the model was conducted based on the following metrics: accuracy (94.73%), precision (93.51%), recall (92.24%), and F1 score (92.87). From the obtained experimental results, it can be concluded that the hybrid system managed to detect fake news successfully.
| Item Type: | Thesis (Other) |
|---|---|
| Subjects: | Q Science > Q Science (General) |
| Divisions: | Faculty of Computing And Information Technology (FACIT) |
| Depositing User: | Cynthia Ugwuoti |
| Date Deposited: | 29 Jul 2026 12:07 |
| Last Modified: | 29 Jul 2026 12:07 |
| URI: | http://eprints.gouni.edu.ng/id/eprint/6071 |
Actions (login required)
 |
View Item |
