A novel approach to fake news classification using LSTM-based deep learning models

• DOI: 10.3389/fdata.2023.1320800
• Publication Date: 2024-01-08T05:05:40Z
• Authors: Halyna Padalko; Vasyl Chomko; Dmytro Chumachenko
• Keywords: Overfitting; Misinformation

The rapid dissemination of information has been accompanied by the proliferation fake news, posing significant challenges in discerning authentic news from fabricated narratives. This study addresses urgent need for effective detection mechanisms. spread on digital platforms necessitated development sophisticated tools accurate and classification. Deep learning models, particularly Bi-LSTM attention-based architectures, have shown promise tackling this issue. research utilized integrating an attention mechanism to assess significance different parts input data. models were trained 80% subset data tested remaining 20%, employing comprehensive evaluation metrics including Recall, Precision, F1-Score, Accuracy, Loss. Comparative analysis with existing revealed superior efficacy proposed architectures. model demonstrated remarkable proficiency, outperforming other terms accuracy (97.66%) key metrics. highlighted potential advanced deep techniques detection. set new standards field, offering combating misinformation. Limitations such as dependency, overfitting, language context specificity acknowledged. underscores importance leveraging cutting-edge methodologies, mechanisms, identification. innovative presented pave way more robust solutions counter misinformation, thereby preserving veracity information. Future should focus enhancing diversity, efficiency, applicability across various languages contexts.