Classification of Indonesian Sarcasm Tweets on X Platform Using Deep Learning

X is a social media platform used to express opinions and criticisms through writing. Many individuals and companies rely on the views or feelings of the general public in making decisions. Consumers tend to trust the opinions of fellow consumers regarding a product or service seen on social media platforms. By monitoring activity on social media, companies that provide products and services can understand the sentiment that consumers feel towards the products they offer. However, since text cannot convey non-verbal cues such as gestures, facial expressions, and intonation, implicit cues such as sarcasm are often difficult to detect. The presence of sarcasm in a tweet can cause the sentiment assessment of the tweet to be inaccurate. Therefore, it is important to perform sarcasm detection, which can significantly improve sentiment analysis results. This study evaluates the performance of four models based on deep learning methods, namely IndoBERT, RoBERTa, BERT base and BERT Multilingual, in recognizing sarcasm in Indonesian on X platform. The experimental results show that the specifically adapted IndoBERT achieves a high F1-score of 95% in dataset that we obtain from previous research that contain various tweet.