Malay Sarcasm Detection on Social Media: A Review, Taxonomy, and Future Directions

Sarcasm is one of the biggest obstacles in sentiment analysis because sarcastic patterns typically describe the inverse meaning of a written sentence. It becomes a major impediment, particularly in social media analytics, because sarcasm reduces the accuracy of a sentence’s connotation. Research on sarcasm detection in Malay text social media is still in its early stages. This paper reviews Malay sarcasm detection on social media and some related work in the past. We present key elements of sarcasm detection along with a proposed taxonomy mapped from the literature review. A detailed taxonomy overview of sarcasm types, approaches, dataset, model-based, and features is presented. A click-based feature is suggested to be categorized as a new feature type because it is independent and is a unique feature embedded with social media itself. In addition, the use of social media datasets in detecting sarcasm has also been discussed in detail, including some examples of datasets used by previous researchers, especially in languages other than English. Finally, several challenging issues in detecting sarcasm are identified and reviewed. Future research direction in this area is suggested at the end of this paper.