Implemented a Deep Learning Model which classifies if a news is authentic or fake based on its title. Trained the Model on GoogleColab.
- Trained on GoogleColab.
- Data Cleaning and preprocessing.
- Feature Engineering.
- Data Visualization.
- Text Processing.
- One hot encoded the Sentences and added Padding to be trained using Embedding Layer in Neural Network.
- Trained the model using Sequential Model and Stacked LSTMS along with Embedding and Dense layers.
- Solved overfitting problems by hyperparameter tuning and Dropout Layers and EarlyStopping Callbacks.
- Dumped the required model using JobLib.
- Tested on RealLife custom Data.
https://www.kaggle.com/c/fake-news
- NLTK
- TensorFlow 2
- Keras
- Joblib
- Scikit-Learn
- Seaborn
- Matplotlib
- Pandas
- Numpy
- String
- re
