mEissa2254/arabart-lr-classifier

AraBART + Logistic Regression — Arabic Text Classifier

Pipeline

1. AraBART generates a summary from the input Arabic text
2. TF-IDF vectorizes both the original text and the generated summary
3. Logistic Regression classifies the combined features

Test Results

• Accuracy : 95.92%
• F1-score : 95.91%

Files

File	Description
arabart/	Fine-tuned AraBART model + tokenizer
lr_model.pkl	Trained Logistic Regression
tfidf_text.pkl	TF-IDF vectorizer for original text
tfidf_sum.pkl	TF-IDF vectorizer for AraBART summaries