Automatic News Agency Detection from News Content

Nowadays, the spread of fake news and rumors has become a severe problem due to the growth of technology. Because technology has dramatically increased the speed of information dissemination, incredibly fake and negative news. The destructive influence of misinformation and disinformation on society and individuals is harmful if they are not detected as early as possible. Some factors can accelerate the pace of spreading false information throughout social media. Firstly, everyone can post and share different types of information in all their forms with just a click. Additionally, the source and accuracy of news information might generally not investigate by users before publishing. Therefore, automatic detection of news sources can play a prominent role in detecting fake news. This study devised a context-source-based model for news source detection from the news context. Three deep recurrent neural networks have been placed at the heart of the proposed system and compared to each other. Through web scraping, we collected a remarkable dataset (more than 420000 news items) from news agency websites to evaluate the quality of the proposed model. According to our results, RNN, LSTM, and GRU models achieved %59, %87, and %88 accuracies, respectively.