Layered Model Stacking - Enhancing DDoS Detection Through Advanced Ensemble Machine Learning Techniques

Aug 2, 2026

Distributed Denial of Service (DDoS) attacks remain a major threat to network infrastructure and services, motivating the development of more reliable detection. This paper introduces DDoS Layered Model Stacking (DDoS LMS), an approach that applies advanced ensemble machine learning to strengthen robustness and detection reliability. The evaluation uses a network-traffic dataset containing legitimate and attack flows, and combines Logistic Regression, k-NN, SVM, MLP, and Naive Bayes. The integrated model reaches 0.9872 accuracy, 0.9829 precision, 0.9847 recall, and 0.9837 F1 score, outperforming individual classifiers.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages4

Reading Time10 min

AccessFree to download

Introduction

Related Works

Layered Model Stacking: Enhancing DDoS Detection Through Advanced Ensemble Machine Learning Techniques Aqeel Sahia, b, ∗ , Member, IEEE, Nabeel Mahdy Haddade , Mohammed Diykhf, g , Shahab Abdullad , Kaled Aljburc , Ali Kutfand , Hayder Al-Hraishawih , Senior Member, IEEE a School of Mathematics, Physics and Computing, University of Southern Queensland, Toowoomba, QLD 4350 Australia b College of Engineering, Al-Shatrah University, Thi-Qar 64001, Iraq c TAFE Queensland, Toowoomba, QLD 4350 Australia dUniSQ College, University of Southern Queensland, Toowoomba, QLD 4350 Australia e Education College, University of Misan, Thi-Qar 64001, Iraq f College of Education for Pure Science, University of Ti-Qar, Thi-Qar 64001, Iraq g Information and Communication Technology Research Group, Al-Ayen University, Thi-Qar 64001, Iraq h Interdisciplinary Centre for Security, Reliability and Trust (SnT), University of Luxembourg, Luxembourg ∗ [Corresponding Author: Aqeel.Sahi@UniSQ.edu.au](Corresponding Author: Aqeel.Sahi@UniSQ.edu.au) Abstract—Distributed Denial of Service (DDoS) attacks continue to cause a substantial threat to network infrastructure and services. In this paper, we propose an approach called DDoS Layered Model Stacking (DDoS LMS) to improve DDoS detection accuracy. Our model uses advanced ensemble machinelearning techniques to enhance the robustness and reliability of detection systems. We evaluate our model using a dataset of network traffic, including both legitimate and attack traffic. Multiple machine learning models are employed, such as Logistic Regression, k-nearest Neighbors (k-NN), Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), and Naive Bayes. Our proposed model, which combines the strengths of these individual classifiers, achieves exceptional results with 0.9872 accuracy, 0.9829 precision, 0.9847 recall, and 0.9837 F1 score. The DDoS LMS notably outperforms individual models and provesits efficiency in detecting DDoS attacks. Index Terms—DDoS, Layered Model Stacking, Ensemble Machine Learning Techniques I. INTRODUCTION Distributed Denial of Service (DDoS) attacks are a primary concern in Internet security [1] . The DDoS attack is a spiteful attempt to interfere with the normal operation of targeted appliances such as servers, services, and networks by overwhelming it with a flood of internet traffic. DDoS attacks have common types consisting of application layer attacks, protocol attacks, and volume-based attacks each affecting distinct aspects of a network or service [2] . These attacks aim to disrupt the normal functioning of targeted systems by inundating them with an overwhelming volume of malicious traffic, rendering them inaccessible to legitimate users. This includes reputation damage and creating a negative user experience. Additionally, there is a major economic loss when businesses experience downtime on their online services or websites. Consequently, DDoS poses a threat to business and critical infrastructure. The most obvious indicator, when a DDoS attack hits, is a service becoming completely unavailable or a site running exceptionally slow. A network connection on the Internet is composed of various components or layers, each serving a distinct purpose. Further, machine learning has an important role in enhancing cybersecurity measures; this paper presents an approach named Layered Model Stacking (DDoS LMS) to improve DDoS detection accuracy and boost the reliability of DDoS detection systems. The DDoS LMS aims to promote the robustness of existing detection mechanisms and hence reinforce network defence against cyber threats. In this context, ensemble machine learning techniques combine multiple learning models to enhance prediction performance, often achieving better results than individual models [3] . By applying ensemble techniques, different classifier architectures are used, such as Logistic Regression, SVM, k-NN, Nave Bayes, and MLP. These different app

FAQs

What problem does the paper address?

The paper addresses the ongoing threat of Distributed Denial of Service (DDoS) attacks to network infrastructure and online services.

How does the proposed DDoS LMS improve DDoS detection?

DDoS LMS uses layered model stacking with ensemble machine learning to combine multiple classifier strengths for more accurate and robust predictions.

Which machine learning models are combined in DDoS LMS?

The approach combines Logistic Regression, k-nearest Neighbors (k-NN), Support Vector Machine (SVM), Multi-Layer Perceptron (MLP), and Naive Bayes.

Aug 2, 2026

Upload by Evangeline