Stance Classification for Fake News Detection with Machine Learning

Abstract

The variety of resources and applications available nowadays made the growth of the news rapid; that allowed people to share their opinions, articles, news, etc.; regardless of the truth percentage they have, which caused the belief that lots of the news can be posted or published through social media and news platforms by an automatic pot or fake user for this purpose. Fake news detection (FND) is a binary classification task; that indicates if the news is right or not right, which involves predicting the probability that a certain news article is designed to be deceptive. Commonly, fake news is produced for political and financial purposes, e.g., influencing presidential elections or manipulating the stock market. Although many studies have been conducted to detect news in English as fake news, the evaluation of the credibility of news written in Arabic is still in its early stage. where FND in Arabic languages got underway to receive more interest in the last years, and many detection approaches present some ability to detect fake news on multiple datasets. Then interest in effective detection models has been growing; specifically, in the Arabic language which has lagged behind the work in other languages. In this paper, we used deep learning models and applied a convolutional neural network and long short-term memory (CNN-BiLSTM) with optimization of Stochastic gradient descent (SDG); to the Arabic accessible dataset called AFND; referring to Arabic Fake News Detection. Our experimental results based on the existing AFND dataset indicate an encouraging and good performance; as we reach an accuracy of 87.7%. We appraise the problem of detecting fake news as one of the classification problems; i.e., our target is to classify a given news as credible or not credible; where credibility is often defined in the sense of believability and quality.

Keywords

• Fake news detection
• Deep learning
• Bidirectional long short-term memory
• Convolutional neural network

References

1. Abdelhady, N., Soliman, T. H. A., & Farghally, M. F. (2022). Stacked-CNN-BiLSTM-COVID: An effective stacked ensemble deep learning framework for sentiment analysis of arabic COVID-19 tweets. https://doi.org/10.21203/rs.3.rs-2206648/v1
2. Alharbi, O. (2021). A deep learning approach combining CNN and Bi-LSTM with SVM classifier for Arabic sentiment analysis. International Journal of Advanced Computer Science and Applications, 12(6),165-172.
3. Al-Yahya, M., Al-Khalifa, H., Al-Baity, H., AlSaeed, D., & Essam, A. (2021). Arabic fake news detection: comparative study of neural networks and transformer-based approaches. Complexity, 2021, 1-10.
4. Baniata, L. H., & Park, S. B. (2016). Sentence representation network for Arabic sentiment analysis. 한국정보과학회 학술발표논문집, 470-472.
5. Ashwaq K., Moath J., Monther A., & Manar J. (2022). AFND: Arabic fake news dataset for the detection and classification of articles credibility. Data in Brief,42, 108141
6. Davoudi, M., Moosavi, M. R., & Sadreddini, M. H. (2022). DSS: A hybrid deep model for fake news detection using propagation tree and stance network. Expert Systems with Applications, 198, 116635.
7. Gauthier -Melancon, G., Ayala, O. M., Brin, L., Tyler, C., Branchaud- Charron, F., Marinier, J., Grande, K., & Le, D. (2022). Azimuth: Systematic error analysis for text classification. Association for Computational Linguistics, 298-310.