intrusion-detection-kaggle
PublicEnd-to-end pipeline for multi-class cyber-attack detection using per-flow network features: data profiling, deduplication, skew-correction, outlier treatment, feature engineering, imbalance handling, and tree-based modeling (XGBoost, LightGBM, CatBoost, stacking), with a final Kaggle submission scoring 0.9146 public / 0.9163 private.
catboostcyber-securitydata-preprocessingensemble-learningfeature-engineeringimbalanced-datajupyter-notebookskagglelightgbmmachine-learning
Creat：2025-04-13T01:33:09
Update：2025-06-22T17:16:33
https://www.kaggle.com/competitions/csai-253-project-phase-2/
0
Stars
0
Stars Increase