A Critique on the Use of Machine Learning on Public Datasets for Intrusion Detection

Catillo, Marta; Vecchio, Andrea Del; Pecchia, Antonio; Villano, Umberto

doi:10.1007/978-3-030-85347-1_19

Cited by 4 publications

(2 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Regarding public datasets, a critical problem was documented by Catillo et al [240]. Despite the rapid growth in the number of articles in this area, many follow a common pattern: propose intrusion detection systems, conduct tests with public datasets, and obtain exceptional detection performance.…”

Section: Discussionmentioning

confidence: 99%

A Comprehensive Survey on Ensemble Learning-Based Intrusion Detection Approaches in Computer Networks

Lucas,

de Figueiredo,

Tojeiro

et al. 2023

IEEE Access

View full text Add to dashboard Cite

Machine learning algorithms present a robust alternative for building Intrusion Detection Systems due to their ability to recognize attacks in computer network traffic by recognizing patterns in large amounts of data. Typically, classifiers are trained for this task. Together, ensemble learning algorithms have increased the performance of these detectors, reducing classification errors and allowing computer networks to be more protected. This research presents a comprehensive Systematic Review of the Literature where works related to intrusion detection with ensemble learning were obtained from the most relevant scientific bases. We offer 188 works, several compilations of datasets, classifiers, and ensemble algorithms, and document the experiments that stood out in their performance. A characteristic of this research is its originality. We found two surveys in the literature specifically focusing on the relationship between ensemble techniques and intrusion detection. We present for the last eight years covered by this survey a timeline-based view of the works studied to highlight evolutions and trends. The results obtained by our survey show a growing area, with excellent results in detecting attacks but with needs for improvement in pruning for choosing classifiers, which makes this work unprecedented for this context.

show abstract

Section: Discussionmentioning

confidence: 99%

A Comprehensive Survey on Ensemble Learning-Based Intrusion Detection Approaches in Computer Networks

Lucas,

de Figueiredo,

Tojeiro

et al. 2023

IEEE Access

View full text Add to dashboard Cite

show abstract

“…Hence, it is advisable to complement evaluations with additional real-world data and remain cautious of potential biases or limitations inherent to any dataset [56]. The CICIDS 2017 dataset, along with other datasets and methodologies, contributes to advancing the development of robust and effective intrusion detection techniques to protect network infrastructures from emerging cyber threats [57].…”

Section: Cicids 2017mentioning

confidence: 99%

From Bytes to Insights: A Systematic Literature Review on Unraveling IDS Datasets for Enhanced Cybersecurity Understanding

Khanan,

Abdelgadir Mohamed,

Mohamed

et al. 2024

IEEE Access

View full text Add to dashboard Cite

In the wake of the expanding digital realm, the imperative for robust cybersecurity measures has burgeoned significantly. This extensive investigation digs into the complicated realm of cybersecurity datasets, with the goal of improving our understanding and implementation of these critical tools. This study's comprehensive evaluation of 37 distinct datasets shows a complicated world in which no one dataset stands out as totally suitable for all uses. A precise balance must be struck between crucial dataset qualities such as diversity, authenticity, and usefulness. Using a complete assessment technique, this paper illuminates the challenges and possibilities that developers and researchers face in the field of cybersecurity datasets. Although some databases accurately identify certain forms of cyberattacks, their coverage may not include the whole range of cyber threats. On the other hand, datasets with a strong emphasis on accurate portrayal may forgo comprehensiveness or practical use. This intricacy is heightened by the dynamic and sophisticated nature of cyber threats, emphasizing the delicate balance required between accuracy and practicality. The study emphasizes the necessity of selecting datasets strategically and contextually for cybersecurity studies, with the goal of matching research objectives with the most appropriate dataset selections. Furthermore, it emphasizes the need of continual cooperation and innovation within the cybersecurity community in developing datasets that accurately represent the ever-changing nature of cyber threats. After analyzing 37 cybersecurity datasets, it is obvious that no one dataset can meet all of the field's unique demands, demonstrating the need of a flexible, adaptable, and developing dataset for intrusion detection systems (IDS). This inquiry offers a critical assessment of dataset characteristics and their related issues, providing essential insights for academics, professionals, and dataset creators, enabling the construction of a more resilient and adaptable cybersecurity infrastructure.

show abstract

Transferability of machine learning models learned from public intrusion detection datasets: the CICIDS2017 case study

et al. 2022

View full text Add to dashboard Cite

A Critique on the Use of Machine Learning on Public Datasets for Intrusion Detection

Cited by 4 publications

References 22 publications

A Comprehensive Survey on Ensemble Learning-Based Intrusion Detection Approaches in Computer Networks

A Comprehensive Survey on Ensemble Learning-Based Intrusion Detection Approaches in Computer Networks

From Bytes to Insights: A Systematic Literature Review on Unraveling IDS Datasets for Enhanced Cybersecurity Understanding

Transferability of machine learning models learned from public intrusion detection datasets: the CICIDS2017 case study

Contact Info

Product

Resources

About