A Review of Machine Learning and Feature Selection Techniques for Cybersecurity Attack Detection with a Focus on DDoS Attacks

Aug 5, 2026

Upload by Ezra

A systematic review examines machine learning methods used in intrusion detection systems, with emphasis on Random Forest, Support Vector Machine, and Decision Tree for identifying Distributed Denial of Service (DDoS) attacks. Following PRISMA, searches across relevant databases retrieved 205 articles, with 68 selected for detailed analysis. Results indicate Random Forest delivers the strongest performance, reaching accuracy up to 99.72% for DDoS detection. Feature selection improves SVM and helps mitigate overfitting in Decision Trees. Ensemble and hybrid approaches further enhance detection accuracy and real-time performance, strengthening modern network cybersecurity.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages17

Reading Time43 min

AccessFree to download

Introduction

Background and Need for IDS

Role of Machine Learning in Cybersecurity

Feature Selection for IDS Performance

RESEARCH ARTICLE OPEN ACCESS A REVIEW OF MACHINE LEARNING AND FEATURE SELECTION TECHNIQUES FOR CYBERSECURITY ATTACK DETECTION WITH A FOCUS ON DDOS ATTACKS 1 Ms Roopesh , 2Nourin Nisha, 3Sasank Rasetti, 4Md Atiqur Rahaman 1 Master of Science in Department of Electrical Engineering, Lamar University, Texas, USA [Email:](Email: muniroopeshraasetti@gmail.com)[ muniroopeshraasetti@gmail.com](Email: muniroopeshraasetti@gmail.com) 2Master of Science in Management Information Systems, College of Business, Lamar University, Texas, USA [Gmail:](Gmail: nishatnitu203@gmail.com)[ nishatnitu203@gmail.com](Gmail: nishatnitu203@gmail.com) 3Master of Science in Department of Electrical Engineering, Lamar University, Texas, USA Email: [srasetti@lamar.edu](srasetti@lamar.edu) 4Department of Management and Information Technology, St. Francis College, New York, USA [Email:](Email: mrahaman4@sfc.edu)[ mrahaman4@sfc.edu](Email: mrahaman4@sfc.edu) This study provides a systematic review of machine learning (ML) techniques applied in intrusion detection systems (IDS), with a particular focus on Random Forest (RF), Support Vector Machine (SVM), and Decision Tree (DT). Following the PRISMA guidelines, a comprehensive search ofrelevant databases identified 205 articles, from which 68 were selected for detailed analysis. The findings highlight that RF consistently outperforms other models, achieving accuracy rates as high as 99.72% in detecting Distributed Denial of Service (DDoS) attacks due to its ensemble learning approach. SVM, while effective in specific scenarios with binary classification tasks, struggles with scalability and high-dimensional datasets, though featureselection significantly improves its performance. DT models, known for their simplicity and interpretability, are prone to overfitting, but this issue is mitigated when combined with featureselection techniques. The study further emphasizes the importance offeature selection in enhancing IDS accuracy and efficiency across various models. Additionally, ensemble and hybrid methods, which combine multiple ML techniques, offer promising improvements in detection accuracy and real-time performance. These findings underscore the potential of machine learning, particularly through the use of ensemble and hybrid approaches, to significantly improve cybersecurity measures in modern networks. KEYWORDS Cybersecurity, Intrusion Detection, Machine Learning, DDoS Attacks, Feature Selection Techniques Submitted: August 12, 2024 Accepted: September 20, 2024 Published: September 22, 2024 Corresponding Author: Ms Roopesh Master of Science in Department of Electrical Engineering, Lamar University, Texas, USA [email:](email: muniroopeshraasetti@gmail.com)[ muniroopeshraasetti@gmail.com](email: muniroopeshraasetti@gmail.com) 10.69593/ajsteme.v4i03 .105 1 Introduction Cybersecurity has become an increasingly critical concern in the digital age, as the growing dependence on internet-based systems and the proliferation of Internet of Things (IoT) devices expose individuals, organizations, and governments to cyber-attacks (Ngo et al., 2023) . Cyber-attacks, such as Distributed Denial of Service (DDoS) attacks, can cause significant damage, leading to service disruptions and financial losses. According to Shrestha et al. (2020), the advent of sophisticated cyber-attacks has outpaced traditional security measures, making it essential to develop advanced tools to detect and mitigate these threats. In this context, Intrusion Detection Systems (IDS) have been the backbone of cybersecurity efforts, aiming to detect unauthorized access and potential security breaches in real-time (Ma et al., 2020) . However, traditional IDS systems often struggle to cope with the increasing complexity and volume of attacks, particularly in the face of distributed and large-scale attacks like DDoS (Mell et al., 2022) . Thus, the use of Machine Learning (ML) techniques has emerged as a

FAQs

What intrusion detection focus and machine learning models does the review analyze?

The review focuses on intrusion detection for DDoS attacks and analyzes Random Forest (RF), Support Vector Machine (SVM), and Decision Tree (DT) approaches.

How does the review select articles, and what is the final sample size?

Using PRISMA guidelines, the study searches relevant databases and screens 205 articles, selecting 68 for detailed analysis.

What impact does feature selection have on model performance in DDoS detection?

Feature selection improves IDS accuracy and efficiency by reducing dimensionality and emphasizing relevant attributes; it also enhances SVM performance and reduces Decision Tree overfitting.

Aug 5, 2026

Upload by Ezra