Student or educator? Sign up with your school email and get 6 months of Pro free.

The DOI matches its Crossref record; 50 of the 50 references it cites were found in a registry or an index; none of the 42 sources its claims cite is checked for support yet.

A DDoS Detection Method Based on Feature Engineering and Machine Learning in Software-Defined Networks

Open access

Abstract

Distributed denial-of-service (DDoS) attacks pose a significant cybersecurity threat to software-defined networks (SDNs). This paper proposes a feature-engineering- and machine-learning-based approach to detect DDoS attacks in SDNs. First, the CSE-CIC-IDS2018 dataset was cleaned and normalized, and the optimal feature subset was found using an improved binary grey wolf optimization algorithm. Next, the optimal feature subset was trained and tested in Random Forest (RF), Support Vector Machine (SVM), K-Nearest Neighbor (k-NN), Decision Tree, and XGBoost machine learning algorithms, from which the best classifier was selected for DDoS attack detection and deployed in the SDN controller. The results show that RF performs best when compared across several performance metrics (e.g., accuracy, precision, recall, F1 and AUC values). We also explore the comparison between different models and algorithms. The results show that our proposed method performed the best and can effectively detect and identify DDoS attacks in SDNs, providing a new idea and solution for the security of SDNs.

Overview

CoChat Summary

TL;DR

This paper proposes an SDN DDoS detection method that combines improved binary grey wolf optimization feature selection with machine learning, selecting 26 of 79 CSE-CIC-IDS2018 features and deploying the best classifier in the SDN controller. Random Forest performed best overall, achieving accuracy 0.9913, precision 0.9843, recall 0.9992, and F1 0.9913 after feature extraction.

Key findings

  • Feature selection reduced the dataset from 79 to 26 features; after feature extraction, all classifiers improved in all metrics, with Decision Tree improving most by 0.0341, followed by RF by 0.0278, and SVM least by 0.0202.
  • On the original dataset, XGBoost had the best accuracy at 0.969. After feature extraction, the conclusion reports RF as best under accuracy, precision, recall, and F1 with 0.9913, 0.9843, 0.9992, and 0.9913, respectively; earlier text notes RF accuracy was slightly lower than Decision Tree but all other metrics were better.
  • AUC-ROC values were SVM 0.9696, RF 0.9917, k-NN 0.9891, DTC 0.9898, and XGBoost 0.99, leading the authors to consider RF the best performer.
  • In the SDN test, an hping3 flooding attack from h5 to h2 disrupted h4 and h6; after mitigation (dropping packets corresponding to a specific input port), h4 and h6 could send data to h2 normally.

Methods

  • Data: CSE-CIC-IDS2018 was cleaned by dropping NaN, empty, or infinite values, reducing samples from 215,761 to 215,076 (685 removed); MinMax scaling mapped features to [0,1]; an average of 210,000 samples was selected, including DDoS, DoS, Brute Force, Infiltration, and Bot attack types.
  • Feature selection: an improved binary grey wolf optimization algorithm used Tent chaotic mapping initialization, nonlinear cosine search parameters, and weighted position updating; it selected 26 features from 79, with fitness based on feature count and k-NN classification error rate.
  • Classifiers and evaluation: SVM, RF, Decision Tree, XGBoost, and k-NN were trained and tested with default parameters and 10-fold cross-validation using accuracy, precision, recall, F1, and ROC/AUC; the best classifier was deployed in a Ryu controller, with experiments on Ubuntu 18.04, 16 GB RAM, AMD Ryzen 7 5800 H at 3.20 GHz, Mininet, and a topology of one controller, two switches, and six hosts.

Limitations

  • Stated: comparison with similar works is difficult because they use different datasets and models; the cited works did not show DDoS attack detection performance on SDNs and did not develop mitigation strategies.
  • Stated future work implies current limitations: the authors plan to use advanced feature engineering and deep learning, research mitigation of adversarial attacks for robustness and adaptability, and integrate the method with other network security technologies.
  • Evident from the text: evaluation uses one dataset (CSE-CIC-IDS2018) and a small Mininet topology (one controller, two switches, six hosts); classifiers used default parameters, and adversarial robustness is not tested in the current work.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3390/s23136176

Cites

50 of 50 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 SDN architecture.
Figure 1 SDN architecture.CC BY · PubMed Central
Figure 2 Hierarchy of grey wolves.
Figure 2 Hierarchy of grey wolves.CC BY · PubMed Central
Figure 3 Flow of the proposed work.
Figure 3 Flow of the proposed work.CC BY · PubMed Central
Figure 4 Distribution of data types in the dataset.
Figure 4 Distribution of data types in the dataset.CC BY · PubMed Central
Figure 5 Comparison of improved and standard convergence factors.
Figure 5 Comparison of improved and standard convergence factors.CC BY · PubMed Central
Figure 6 The importance of each feature.
Figure 6 The importance of each feature.CC BY · PubMed Central
Figure 7 Support Vector Machine (SVM).
Figure 7 Support Vector Machine (SVM).CC BY · PubMed Central
Figure 8 Random Forest (RF).
Figure 8 Random Forest (RF).CC BY · PubMed Central
Figure 9 Confusion matrix.
Figure 9 Confusion matrix.CC BY · PubMed Central
Figure 10 Performance indicators of each classifier in the original dataset.
Figure 10 Performance indicators of each classifier in the original dataset.CC BY · PubMed Central
Figure 11 Performance indicators of each classifier in the dataset after feature extraction.
Figure 11 Performance indicators of each classifier in the dataset after feature extraction.CC BY · PubMed Central
Figure 12 The confusion matrix of each classifier on the original dataset.
Figure 12 The confusion matrix of each classifier on the original dataset.CC BY · PubMed Central
Figure 13 The confusion matrix of each classifier on the dataset after feature extraction.
Figure 13 The confusion matrix of each classifier on the dataset after feature extraction.CC BY · PubMed Central
Figure 14 AUC-ROC curves for all classifiers.
Figure 14 AUC-ROC curves for all classifiers.CC BY · PubMed Central
Figure 15 Network topology.
Figure 15 Network topology.CC BY · PubMed Central
Figure 16 Response of the controller to DDoS attacks.
Figure 16 Response of the controller to DDoS attacks.CC BY · PubMed Central
Figure 17 Mitigating controller response after DDoS attacks.
Figure 17 Mitigating controller response after DDoS attacks.CC BY · PubMed Central

Verification

Claims the paper makes

0 Valid · 1 Not valid · 20 Not checked · 33 Own results and interpretations

12 more claims

Create a free account to see all 54 claims and to verify the 42 references still open.

See all claims — free account
  • Probably not a substantial claim. 22
  • It may state two separate claims as one. 25
  • The quote only partly states this claim. 2

Create a free account to see these claims.

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

1 of 1

Citations

Cites

Show all 50 references

Cited by

See all 119

Metadata

Authors
Liu, Z., Wang, Y., Feng, F., Liu, Y., Li, Z., & Shan, Y.
Affiliations
Hebei University (CN)
Corresponding author
Liu, Yifan
Venue
Sensors
Publisher
MDPI AG
Published
2023-07-05
Volume
23
Issue
13
Pages
6176
ISSN
1424-8220 (ISSN-L)
OpenAlex
W4383346073
Cited by
119
References
50
Citations by year
2026: 43 · 2025: 38 · 2024: 34 · 2023: 4
FWCI
9.74
Type
journal article
Language
en
Open access
Locations
  • Sensors (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • National Natural Science Foundation of Hebei Province, China (F2019201427, 2017A20004)
  • Fund for Integration of Cloud Computing and Big Data, Innovation of Science and Education (FII) of Ministry of Education of China (F2019201427, 2017A20004)
Topics
  • Physical Sciences › Computer Science › Computer Networks and Communications › Network Security and Intrusion Detection
  • Physical Sciences › Computer Science › Computer Networks and Communications › Software-Defined Networks and 5G
  • Physical Sciences › Computer Science › Hardware and Architecture › Network Packet Processing and Optimization
Keywords
DDoS attack detection, software-defined networking, feature selection, grey wolf optimizer, random forest, XGBoost, support vector machine, k-nearest neighbors, decision tree, SDN controller security, network intrusion detection
MeSH terms
Machine Learning, Algorithms, Software, Cluster Analysis, Benchmarking

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

A language model drew the claims from the paper on 2026-11-16, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-11-16 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: A DDoS Detection Method Based on Feature Engineering and Machine Learning in Software-Defined Networks. https://app.cochat.ai/papers/10.3390/s23136176
The paper
Liu, Z., Wang, Y., Feng, F., Liu, Y., Li, Z., & Shan, Y. (2023). A DDoS Detection Method Based on Feature Engineering and Machine Learning in Software-Defined Networks. Sensors. https://doi.org/10.3390/s23136176

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.