Binary Classifier: Spam Detection

This project uses a Naive Bayes classifier to detect spam messages in SMS data.

Dataset

• Source: SMSSpamCollection
• 5572 messages with 2 columns: label and message

Approach

1. Load and clean the data (remove missing values and duplicates)
2. Map labels (ham → 0, spam → 1)
3. Vectorize text using CountVectorizer
4. Train a Multinomial Naive Bayes classifier
5. Evaluate with accuracy and a confusion matrix

Results

• Accuracy: ≈ 97.6%
• Confusion Matrix:
• Test Prediction Example: "Congratulations! You won a free gift card." → Spam (1)

Tools Used

• Python, Pandas, Scikit-learn, Matplotlib, Seaborn, Jupyter Notebook