As part of my Troy Tech internship under Joshua Tallman at Concordia University, I built anomaly detection pipelines on the NF-UNSW-NB15 dataset using Random Forest.
Standardized and scaled 1.6M NetFlow V1 records for reliable training
Subnet frequency aggregations and numeric transformations
Baseline Random Forests trained across multiple feature sets
Precision, recall, F1-score, ROC-AUC, and cross-validation