5 Arabic news classification Models using TensorFlow. Takes user input through a GUI.
Models include + Credits:
- CNN - Youssef Farag
- SVM - Youssef Farag
- HAN - Youssef Farag
- KNN - Youssef Farag
- Decision Tree - Mohab Abdulatif
- LSTM - Ahmed Hossam
- KNN - Ahmed Diaa
Evaluations:
- CNN -> 0.98 accuracy
- SVM -> 0.81 accuracy
- HAN -> 0.92 accuracy
- KNN -> 0.98 accuracy
- Decision Tree -> 0.93 accuracy
- LSTM -> 0.91 accuracy
- KNN -> 0.94 accuracy
Classes include:
- Politics
- Entertainment
- Economy
- Sports
- First download the dataset.
- Run the main GUI file after updating paths.
- You will have the option to train or classify.
*Make sure the generated x_model.h5 and word2vec.model path is hooked to the classify file.
OLD GUI:
New GUI: