Network Attack Detection using Machine Learning

Developed a machine learning–based intrusion detection system for IoT environments using a 1.5M-record subset of the Bot-IoT dataset, targeting DoS, DDoS, and Reconnaissance attacks.

Applied comprehensive preprocessing (missing values, encoding, log transformation, scaling, rare category handling, feature selection, and PCA) to enhance data quality and model performance.

Trained a Random Forest classifier achieving strong results on ~300K test samples using Python, Pandas, NumPy, Scikit-learn, Matplotlib, and Seabor