Sarcasm_Detection

URL: https://github.com/Akshitaag/Sarcasm_Detection

Description: No description available.

Related Paper

This repository contains the code for the paper:
“Did you really mean what you said? Sarcasm Detection in Hindi-English Code-Mixed Data using Bilingual Word Embeddings”

• Paper link: ACL Anthology

Abstract

With the increasing use of social media, many new NLP challenges have emerged, including sarcasm detection. This study presents:

• A corpus of tweets for training custom word embeddings.
• A Hinglish (Hindi-English code-mixed) dataset labeled for sarcasm detection.
• A deep learning-based approach leveraging bilingual word embeddings derived from FastText and Word2Vec techniques.

Models and Results

The study experimented with various deep learning models:

• CNNs
• LSTMs
• Bi-directional LSTMs (with and without attention)

The attention-based Bi-directional LSTM achieved the best performance, with an accuracy of 78.49%, outperforming previous state-of-the-art approaches.