Impact of Resampling techniques in Deep Learning based Intrusion Detection: A Comparative Study on NSL-KDD and UNSW-NB15
Bibliographic record
Abstract
This study examines the effectiveness of a specific set of data resampling techniques within a deep learning framework to enhance network intrusion detection system (NIDS) which is a critical aspect of cybersecurity. We evaluate a CNN-BLSTM model for attack classification on two widely used benchmark NIDS datasets (NSL-KDD and UNSW-NB15) that exhibit varying degrees of class imbalance which can undermine the detection of rare yet critical cyberattacks. Our approach integrates Adaptive Synthetic (ADASYN) oversampling both independently and in combination with undersampling methods such as Random UnderSampling (RUS), TomekLinks and One-Sided Selection (OSS) to address this imbalance. The results demonstrate that for the NSL-KDD dataset which suffers from severe imbalance, ADASYN coupled with OSS significantly improves classification accuracy. In contrast, for the UNSW-NB15 dataset, with relatively milder imbalances, the deep learning model performs comparably well without extensive resampling. Beyond the standard metrics, we also conduct additional evaluations through class-wise accuracy analysis and permutation-based feature importance to better understand the impact of resampling on minority class performance and feature relevance. These findings underscore the importance of tailoring resampling strategies to the specific class distributions within NIDS datasets to optimize deep learning performance in cybersecurity applications.
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 | 0.005 | 0.013 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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".