Automatic Identification of Fake News Using Deep Learning

• Published in: 2019 Sixth International Conference on Social Networks Analysis, Management and Security (SNAMS) pp. 383 - 388
• Main Authors: Qawasmeh, Ethar, Tawalbeh, Mais, Abdullah, Malak
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.10.2019
• Subjects: Analytical models; Biological system modeling; Data models; deep learning; Fake news challenge; fake news detection; Feature extraction; Machine learning; Neural networks; social media; Social networking (online)
• DOI: 10.1109/SNAMS.2019.8931873

The rapid development of computing trends, wireless communications, and the smart devices industry has contributed to the widespread of the internet. People can access internet services and applications from anywhere in the world at any time. There is no doubt that these technological advances have made our lives easier and saved our time and efforts. On the other side, we should admit that there is a misuse of internet and its applications including online platforms. As an example, online platforms have been involved in spreading fake news all over the world to serve certain purposes (political, economic, or social media). Detecting fake news is considered one of the hard challenges in term of the existing content-based analysis of traditional methods. Recently, the performance of neural network models have outperformed traditional machine learning methods due to the outstanding ability of feature extraction. Still, there is a lack of research work on detecting fake news in news and time critical events. Therefore, in this paper, we have investigated the automatic identification of fake news over online communication platforms. Moreover, We propose an automatic identification of fake news using modern machine learning techniques. The proposed model is a bidirectional LSTM concatenated model that is applied on the FNC-1 dataset with 85.3 % accuracy performance.