Home / Projects / Machine Learning Topic Classifier
Machine Learning Topic Classifier
Description
A machine learning project that classifies news articles into predefined topics using a fine-tuned BERT deep learning model.
Tech
Python, Flask, BERT, scikit-learn, Google Colab, pandas, NumPy, Matplotlib
Key Points
- Automatic news article classification
- Fine-tuned BERT model with 2,000+ labeled samples
- Optimized training for best validation performance
- Comparison with TF-IDF + Naive Bayes baseline
- Flask backend for model inference
- User-friendly web interface
- Gdown for large model file handling
Challenges
Fine-tuning BERT for optimal performance, TF-IDF and Naive Bayes baseline implementation.
Links
Media
