A Simplified Method to Identify the Sarcastic Elements of Bahasa Indonesia in Youtube Comments
Wilbert Wijaya, I Made Murwantara, Aditya Rama Mitra · 2020
Comments on Youtube media provide feedback to the content owner of how their video influence audience. However, some comments may have intention to convey readers not to the original goals of the video content. Moreover, the sarcasm statements may also provide misleading information that possibly put the conversation into a hatred comments. This paper presents a simplified method to identify sarcasm within comments in a youtube video channel, which enable the channel or video owner to identify and remove or eliminate some comments that may arise hatred between the audience. To identify sarcasm comments, we use Naïve Bayes to find out which statements may have sarcasm elements. Our approach, specifically, designed to detect and explicate the elements of sarcastic words or statements in Bahasa Indonesia. To demonstrate our approach, an Indonesian political Youtube Channel comments for the range of 7 months prior to the 2019 Indonesian Presidential election has been collected and analysed. Our result shows that Naïve Bayes identified 79% of sarcasm within the comments, with accuracy 0,802, precision 1.0 and recall 0,651.