SMS and Email Spam Classification Using Machine Learning

• I designed an end-to-end spam detection system that processes over 16,000 SMS and email messages. For the NLP operations, Python with Scikit-learn was utilized with TF-IDF and text data preprocessing.
• Three machine learning classifiers were implemented: Linear SVM, Multinomial Naive Bayes, and XGBoost. After hyperparameter tuning and decision threshold optimization, the best accuracy of 86.57% was obtained with XGBoost.
• I also designed a prediction and evaluation system using Pandas, NumPy, and Flask/Shiny integration. This increased classification accuracy by using an F1 score of 0.655 and reducing false positives.