Baseline experiments on training a Decision Tree Classifier and a Random Forest Classifier using Grid Search with Cross Validation on the CIC IDS 2018 dataset for training Machine Learning network intrusion detection classifier models.

Machine Learning Based - Intrusion Detection System

En este proyecto se evalúan y comparan diferentes técnicas de aprendizaje automático para la detección de intrusiones en red.

IoT intrusion detection project enhancing detection accuracy of DoS/DDoS attacks through data imbalance correction using SMOTE and machine learning classifiers, achieving an F1-score of 0.999983.

Machine learning based network intrusion detection system

Addressing Class Imbalance in CIC-IDS-18. Improve intrusion detection accuracy and reduce false alarms by tackling class imbalance. Utilizing artificial oversampling techniques and comparing their efficacy with deep neural network algorithms.Tech: Python, Jupyter-Notebook, Scik

Unsupervised network intrusion detection on CIC-IDS-2018 using TRCC density-peak clustering + SEDE hyperparameter optimization. Chunked architecture (1k-row micro-batches), MCC-optimized, with per-dataset density-inversion mapping for volumetric attacks. The only model in the comparison with above-random MCC on Bot/DDoS traffic.