
Project Overview
Prepared text corpus, applied TF-IDF vectorization and trained Multinomial Naive Bayes to classify emails; prioritized high precision to reduce false positives.
Tools & Techniques
PythonNLTKNaive BayesScikit-learn
Key Outcomes
- Achieved >97% accuracy on public datasets.
- Low false-positive rate after threshold tuning.
- Lightweight model suitable for production filtering.