An empirical study of pattern leakage impact during data preprocessing on machine learning-based intrusion detection models reliability

In this paper, we investigate the impact of pattern leakage during data preprocessing on the reliability of Machine Learning (ML) based intrusion detection systems (IDS). Data leakage, also known as pattern leakage, occurs during data preprocessing when information from the testing set is used in tr...

Full description

Bibliographic Details
Main Authors: Bouke, Mohamed Aly, Abdullah, Azizol
Format: Article
Published: Elsevier B.V. 2023
Online Access:http://psasir.upm.edu.my/id/eprint/106552/
_version_ 1848864782027849728
author Bouke, Mohamed Aly
Abdullah, Azizol
author_facet Bouke, Mohamed Aly
Abdullah, Azizol
author_sort Bouke, Mohamed Aly
building UPM Institutional Repository
collection Online Access
description In this paper, we investigate the impact of pattern leakage during data preprocessing on the reliability of Machine Learning (ML) based intrusion detection systems (IDS). Data leakage, also known as pattern leakage, occurs during data preprocessing when information from the testing set is used in training, leading to overfitting and inflated accuracy scores. Our study uses three well-known intrusion detection datasets: NSL-KDD, UNSW-NB15, and KDDCUP99. We preprocess the data to create versions with and without pattern leakage and train and test six ML models: Decision Tree (DT), Gradient Boosting (GB), K-neighbours (KNN), Support Vector Machine (SVM), Random Forest (RF), Logistic Regression (LR). Our results show that building IDS models with data leakage leads to higher accuracy but is unreliable. Additionally, we find that some algorithms are more sensitive to data leakage than others, as seen by the drop in model accuracy when built without leakage. To address this problem, we provide suggestions for mitigating data leakage in the training process and analyzing the sensitivity of different algorithms. Overall, our study emphasizes the importance of addressing data leakage in the training process to ensure the reliability of ML-based IDS models.
first_indexed 2025-11-15T13:54:17Z
format Article
id upm-106552
institution Universiti Putra Malaysia
institution_category Local University
last_indexed 2025-11-15T13:54:17Z
publishDate 2023
publisher Elsevier B.V.
recordtype eprints
repository_type Digital Repository
spelling upm-1065522024-10-03T04:25:37Z http://psasir.upm.edu.my/id/eprint/106552/ An empirical study of pattern leakage impact during data preprocessing on machine learning-based intrusion detection models reliability Bouke, Mohamed Aly Abdullah, Azizol In this paper, we investigate the impact of pattern leakage during data preprocessing on the reliability of Machine Learning (ML) based intrusion detection systems (IDS). Data leakage, also known as pattern leakage, occurs during data preprocessing when information from the testing set is used in training, leading to overfitting and inflated accuracy scores. Our study uses three well-known intrusion detection datasets: NSL-KDD, UNSW-NB15, and KDDCUP99. We preprocess the data to create versions with and without pattern leakage and train and test six ML models: Decision Tree (DT), Gradient Boosting (GB), K-neighbours (KNN), Support Vector Machine (SVM), Random Forest (RF), Logistic Regression (LR). Our results show that building IDS models with data leakage leads to higher accuracy but is unreliable. Additionally, we find that some algorithms are more sensitive to data leakage than others, as seen by the drop in model accuracy when built without leakage. To address this problem, we provide suggestions for mitigating data leakage in the training process and analyzing the sensitivity of different algorithms. Overall, our study emphasizes the importance of addressing data leakage in the training process to ensure the reliability of ML-based IDS models. Elsevier B.V. 2023 Article PeerReviewed Bouke, Mohamed Aly and Abdullah, Azizol (2023) An empirical study of pattern leakage impact during data preprocessing on machine learning-based intrusion detection models reliability. Expert Systems with Applications, 230. pp. 1-9. ISSN 0957-4174; ESSN: 1873-6793 https://linkinghub.elsevier.com/retrieve/pii/S0957417423012174 10.1016/j.eswa.2023.120715
spellingShingle Bouke, Mohamed Aly
Abdullah, Azizol
An empirical study of pattern leakage impact during data preprocessing on machine learning-based intrusion detection models reliability
title An empirical study of pattern leakage impact during data preprocessing on machine learning-based intrusion detection models reliability
title_full An empirical study of pattern leakage impact during data preprocessing on machine learning-based intrusion detection models reliability
title_fullStr An empirical study of pattern leakage impact during data preprocessing on machine learning-based intrusion detection models reliability
title_full_unstemmed An empirical study of pattern leakage impact during data preprocessing on machine learning-based intrusion detection models reliability
title_short An empirical study of pattern leakage impact during data preprocessing on machine learning-based intrusion detection models reliability
title_sort empirical study of pattern leakage impact during data preprocessing on machine learning-based intrusion detection models reliability
url http://psasir.upm.edu.my/id/eprint/106552/
http://psasir.upm.edu.my/id/eprint/106552/
http://psasir.upm.edu.my/id/eprint/106552/