A Novel Machine Learning Model for Sarcasm Detection on Facebook Comments

This research focuses on identifying sarcasm in online comments using a combination of machine learning techniques, including feature extraction, pattern recognition, and analysis of emojis. Various methods were utilized to categorize comments into sarcastic and non-sarcastic categories, aiding computers in understanding the sentiment expressed. The experiments have yielded positive results, particularly with hierarchical clustering, which exhibited higher Silhouette Scores and improved cluster quality compared to other techniques. The study used a dataset containing 5,471 comments, labeled with sarcasm as (0 for sarcastic, 1 for non-sarcastic), with sentiment analysis facilitating comment categorization. By categorizing comments based on sentiment, computers can easily identify the underlying sentiment of each comment. The combination of features, including emojis and text, adds an extra layer of meaning, helping the system to identify the type of comment easily. The objective is to help computers understand human language better, enabling them to grasp the real meaning behind a joke and sarcastic sentence, there by making online communication more accurate and effective.