Masked Statement Classification in Conversational Text

Abstract

Sarcasm can be really tricky for computer programs that analyze text to understand people’s feelings and opinions. This is because sarcasm involves saying something with the opposite meaning of what you actually feel, often to make fun of a situation. When computers analyze text, they focus on the words used rather than the intended meaning, which can lead to mistakes in how the text is categorized. Sarcasm also creates problems when trying to figure out the author’s true emotions in tasks like sentiment analysis. So, there is a need of a way to identify sarcasm in text and correct the sentiment it conveys. Recent studies have employed deep learning and machine learning methods to identify sarcasm. Deep learning is quite powerful, but traditional machine learning methods are still widely used and can work as well if trained properly. This study aims to compare different algorithms like Long Short Term Memory (LSTM), K-Nearest Neighbors (KNN), Computational Neural Network(CNN), and Artificial Neural Network(ANN) etc to see how well they work in sarcasm detection. The survey examined several systems and algorithms for detecting sarcasm, as well as any faults or drawbacks associated with these approaches. It also discusses various important challenges and future directions for sarcasm detection research in order to provide useful insights in this field