Drift Robust Evaluation of Machine Learning Intrusion Detection Methods on CSE-CIC-IDS2028 Dataset
Bibliographic record
Abstract
This study introduces a deployment-oriented, time-based evaluation framework for assessing drift robustness in machine learning-based network intrusion detection systems (NIDS) using the CSE-CIC-IDS2018 dataset. To replicate real-world deployment, where models encounter unseen, drifted future data, traffic flows are logically divided into training (days 1–6), validation (days 7–8), and testing (days 9–10) sets. Multiple operational thresholds that correspond to different security operations center (SOC) capacity constraints are used to benchmark six representative models: Random Forest, XGBoost, LightGBM, CatBoost, Logistic Regression, and Passive-Aggressive Classifier. The outcomes show that tree-based ensembles perform significantly better than adaptive and linear methods. With 94.1% precision, 78.0% recall, and only 358 false alarms per hour at a 1% false positive rate budget, LightGBM attains ideal deployment characteristics. Hourly temporal analysis shows that aggregate metrics hide important performance dynamics. For example, tree models keep stable discrimination, but adaptive classifiers get much worse when the distribution shifts. Threshold sensitivity analysis shows that FPR budgets of 0.5–1% strike a balance between detection coverage and analyst workload, while validation-frozen thresholds lead to false alarm rates that are not practical for operations.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.011 | 0.015 |
| Meta-epidemiology (narrow) | 0.003 | 0.000 |
| Meta-epidemiology (broad) | 0.002 | 0.001 |
| Bibliometrics | 0.004 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.002 | 0.001 |
| Research integrity | 0.002 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".