An empirical assessment of ML models for 5G network intrusion detection: A data leakage-free approach

e-Prime - Advances in Electrical Engineering Electronics and Energy, Год журнала: 2024, Номер 8, С. 100590 - 100590

Опубликована: Май 9, 2024

This paper thoroughly compares thirteen unique Machine Learning (ML) models utilized for Intrusion detection systems (IDS) in a meticulously controlled environment. Unlike previous studies, we introduce novel approach that avoids data leakage, enhancing the reliability of our findings. The study draws upon comprehensively labeled 5G-NIDD dataset covering broad spectrum network behaviors, from benign real-user traffic to various attack scenarios. Our preprocessing and experimental design have been carefully structured eradicate any standout feature methodology significantly improves robustness dependability results compared prior studies. ML are evaluated using performance metrics, including accuracy, precision, recall, F1-score, ROC AUC, execution time. reveal K-Nearest Neighbors model is superior accuracy while Voting Classifier stands out precision F1-score. Decision Tree, Bagging, Extra Trees exhibit strong recall scores. In contrast, AdaBoost falls short across all assessed metrics. Despite displaying only modest on other Naive Bayes excels computational efficiency, offering quickest emphasizes importance understanding models' distinct strengths, drawbacks, trade-offs intrusion detection. It highlights no single universally superior, choice hinges nature dataset, specific application requirements, resources available.

Язык: Английский

---

Implications of Data Leakage in Machine Learning Preprocessing: A Multi-Domain Investigation

Research Square (Research Square), Год журнала: 2024, Номер unknown

Опубликована: Июль 12, 2024

Data leakage during machine learning (ML) preprocessing is a critical issue where unintended external information skews the training process, resulting in artificially high-performance metrics and undermining model reliability. This study addresses insufficient exploration of data across diverse ML domains, highlighting necessity comprehensive investigations to ensure robust dependable models real-world applications. Significant discrepancies performance due were observed, with notable variations F1 scores ROC AUC values for Breast Cancer dataset. The Tic-Tac-Toe Endgame dataset analysis revealed varying impact on like Ridge, SGD, GaussianNB, MLP, underscoring profound effect leakage. German Credit Scoring showed slight enhancements recall DT GB without leakage, indicating reduced overfitting. Additionally, such as PassiveAggressive, Nearest Centroid exhibited shifts metrics, intricate response also raw rates, 6.79% Spambase 1.99% Cancer. These findings emphasize meticulous management validation mitigate effects, which crucial developing reliable models.

Язык: Английский

---

Advancing Intrusion Detection: Application of Distributed Deep Learning on the KDD Cup 99 Dataset

International Journal of Electronics and Communication Engineering, Год журнала: 2024, Номер 11(6), С. 107 - 113

Опубликована: Июнь 30, 2024

Intrusion Detection Systems (IDS) are crucial for protecting IT infrastructures against increasingly sophisticated and evolving threats. Faced with complex attacks such as stealthy or polymorphic threats, conventional methods based on rules signatures show their limitations. An innovative IDS approach utilizing a deep neural network integrated into distributed architecture dynamic precise traffic analysis is introduced. Tested the KDD Cup 99 dataset, this method demonstrated an accuracy of 99.90%, recall 99.89%, specificity 100%, marking significant improvement over traditional systems. The exceptional performance obtained encourages broader adoption system suggests potential revolutionizing security practices. implications findings current strategies also discussed, directions future research proposed.

Язык: Английский

---

Towards robust and efficient intrusion detection in IoMT: a deep learning approach addressing data leakage and enhancing model generalizability

Multimedia Tools and Applications, Год журнала: 2024, Номер unknown

Опубликована: Июль 30, 2024

Язык: Английский

---

A novel intelligent fault diagnosis method of helical gear with multi-channel information fused images under small samples

Applied Acoustics, Год журнала: 2024, Номер 228, С. 110357 - 110357

Опубликована: Окт. 31, 2024

Язык: Английский

---