The web application uses a pre-trained LSTM model for feature extraction. The extracted features are concatenated with the input data and passed through a Dense neural network for classification.
Python
Flask
TensorFlow
Transformers
Pandas
NumPy
scikit-learn
Dataset
Dataset used in this project was downloaded from kaggle
https://www.kaggle.com/code/therealsampat/fake-news-detection/input
Difference from Existing
As you can see in this kaggle repository they used different machine learning models for this work but in our work we used LSTM for features extraction and those features then fed to a dense neural network for detection.
Architecture Diagram