This project leverages Naïve Bayes (Bernoulli, Multinomial, and Gaussian) with TF-IDF vectorization to classify emails as spam or ham with ~98% precision and ~97% accuracy. 🔹 Key Highlights: - Text preprocessing for high-quality input: cleaning, (null values, stop word, punctuation) removal, EDA, tokenization, etc with the help of NLTK, Seaborn, Matplotlib, Numpy libraries. - Feature extraction with TF-IDF (max features tuning) : Vectorization - Hyperparameter tuning for improved accuracy using GridSearchCV - Model comparison & evaluation (precision, confusion matrix) using ScikitLearn - Deployed using Streamlit for a user-friendly experience