Using Linguistic and Stylometric Features for Satirical News Detection in Bahasa Indonesia

In this paper, we present our work on satirical news detection for Indonesian. Despite the progress of satirical news detection in English, there has been lack of research work for other languages including Indonesian. We start this research by developing INSIDE, a novel dataset for Indonesian Satirical News Detection. The dataset consists of 100 satirical news and 190 its corresponding real news articles that were collected from satirical source and various legitimate online sites. INSIDE is one of the efforts to encourage more research in low-resource language. It also enrich real-world data source for developing cross-lingual model in the area of satirical news detection. Further to dataset development, we conducted experiment on satirical news detection using various linguistic and stylometric features. Results show the effectiveness of our model. In addition to that, certain features such as bag-of-words, writing style and structural features are found to be beneficial for detecting satirical news.