The Fake News Classifier using Bidirectional LSTM project aims to develop and train a deep learning model to classify news articles as either real or fake. Leveraging the capabilities of Bidirectional Long Short-Term Memory (BiLSTM) networks, this project focuses on accurately analyzing and categorizing news content based on textual data.
Before feeding the textual dataset into the sequence model, extensive preprocessing is required to convert the text data into sequences of vectors. This repository provides a comprehensive, step-by-step guide on the necessary preprocessing steps, model building, training, and evaluation processes.
- Accurate Fake News Detection: Utilizes a Bidirectional LSTM model to distinguish between real and fake news articles.
- Comprehensive Text Analysis: Includes NLP techniques such as tokenization, stopwords removal, and padding to prepare text data for the model.
- Detailed Preprocessing: Demonstrates the preprocessing steps needed to clean and convert text data into a format suitable for deep learning models.
- Model Training and Evaluation: Guides through the process of building, training, and evaluating the performance of the Bidirectional LSTM model.
- Importing fake and real news datasets with Pandas and checking sizes and null values
- Exploring the imported dataset and feature engineering
- Performing data cleaning
- Visualizing the cleaned dataset
- Preparing the dataset for training by performing tokenization and padding
- Build and train a bidirectional LSTM model
- Evaluating the model performance
Contributions are welcome! If you have suggestions or improvements, please fork the repository and submit a pull request. For major changes, please open an issue first to discuss what you would like to change.