Sarcasm Detector on News Headline with EDA and RNN

URL: https://github.com/shamiulshifat/sarcasm-detector-on-news-headline-with-EDA-and-RNN

Description

This project focuses on sarcasm detection in news headlines using a publicly available dataset. The process involves data preprocessing, including:

• Cleaning the text
• Word embedding

Following preprocessing, a Bidirectional LSTM model was developed to predict whether a given news headline is sarcastic or not.

Methods and Models

• Exploratory Data Analysis (EDA) to understand the dataset.
• Data Cleaning & Preprocessing, including tokenization and word embeddings.
• Bidirectional Long Short-Term Memory (BiLSTM) for sarcasm classification.

Dataset

The dataset used in this project contains news headlines labeled as sarcastic or non-sarcastic.

Implementation Details

The implementation includes:

• Data preprocessing (cleaning and embedding)
• Feature engineering
• Model training using BiLSTM
• Evaluation of model performance

Requirements

The project is implemented in Python using deep learning libraries.

Results

The model’s performance is evaluated based on classification accuracy, but specific metrics are not provided in the repository.