Published Paper Details:

Sarcasm detection is a vital component of natural language processing with relevance in various real-world applications. Sarcasm, characterized by statements that convey the opposite of their literal meaning, can lead to misinterpretation and misunderstanding in textual data. Therefore, the necessity for sarcasm detection arises in applications such as sentiment analysis, content moderation on social media, customer feedback analysis, text summarization, and human-computer interactions, including chatbots and virtual assistants. Detecting sarcasm ensures that sentiment analysis accurately reflects the intended emotions, aids in maintaining respectful online environments, and enhances customer insights.This research paper presents a sarcasm detection model that leverages pre-trained word embeddings from GloVe (Global Vectors for Word Representation) combined with a Long Short-Term Memory (LSTM) neural network architecture. The goal of this study is to develop an effective model for detecting sarcasm in textual data.The heart of the proposed model is the LSTM layer, a type of recurrent neural network that can capture contextual information in sequences. The model is trained on a labeled dataset of text examples, with binary labels indicating the presence or absence of sarcasm. Experimental results demonstrate the model's effectiveness in detecting sarcasm, with a focus on achieving high accuracy and generalization performance. A visualization of the learning process shows the convergence of training and validation metrics over epochs.The combination of pre-trained embeddings and LSTM architecture highlights the potential for improved sarcasm detection in various applications, including social media analysis, sentiment analysis, and content moderation. This research contributes to the ongoing development of natural language understanding models and their application in sarcasm detection.