Content Based Fake News Detection Using N-Gram Models

Fake news is very popular these days because of the increasing popularity of social media. Detecting fake news is considered as one of the most dangerous types of deception because it is created with dishonest intention to misdirect the public. Many researchers proposed fake news detection systems considering many approaches; content, social-context, and propagation. When the news is detected fake or real, there is a limitation in the accuracy and understandability of language. In this paper, we propose the fake news detection system that considers the content of the online news articles. We investigate two machine learning algorithms with the use of word n-grams and character n-grams analysis. Experiments yield better results using character n-grams with Term-Frequency-Inverted Document Frequency (TF-IDF) and Gradient Boosting Classifier achieves an accuracy of 96%.