Automatic fake news detection is a challenging problem in deceptiondetection, and it has tremendous real-world political and social impacts.However, statistical approaches to combating fake news has been dramaticallylimited by the lack of labeled benchmark datasets. In this paper, we presentliar: a new, publicly available dataset for fake news detection. We collected adecade-long, 12.8K manually labeled short statements in various contexts fromPolitiFact.com, which provides detailed analysis report and links to sourcedocuments for each case. This dataset can be used for fact-checking research aswell. Notably, this new dataset is an order of magnitude larger than previouslylargest public fake news datasets of similar type. Empirically, we investigateautomatic fake news detection based on surface-level linguistic patterns. Wehave designed a novel, hybrid convolutional neural network to integratemeta-data with text. We show that this hybrid approach can improve a text-onlydeep learning model.
展开▼
