Sentiment Analysis of Sarcastic Hindi Sentences: Analysis of SVM and XGBoost Models

Detecting sarcasm in text presents a significant challenge due to its tendency to subvert literal sentiment. This study proposes a novel approach for classifying sarcastic sentences as either positive or negative. We developed a machine learning model that integrates contextual and syntactic features to enhance sarcasm detection. A major contribution of this research is the creation of a specialized dataset of sarcastic sentences derived from movie reviews, specifically curated for this study. Our model, trained on this dataset, demonstrates superior performance over traditional sentiment analysis techniques in accurately identifying the underlying sentiment of sarcastic remarks. The enhanced precision of our approach holds substantial implications for applications in movie review analytics, social media analytics, customer feedback systems, and automated content evaluation.