Evaluation of the performance of supervised and unsupervised Machine learning techniques for intrusion detection

Pórtela, Fernando Gutiérrez; Mendoza, Florina Almenares; Benavides, Liliana Calderón

doi:10.1109/icasat48251.2019.9069538

Cited by 9 publications

(7 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Through a methodical removal of less informative information, RFE improves the intrusion detection system's efficiency and interpretability. [31] This approach offers a significant benefit in determining the most important variables for categorization.…”

Section: Random Forest Classifiermentioning

confidence: 99%

GuardianAI: Smart Intrusion Detection for Modern Threats

SAIRAJKUMAR

2024

Preprint

View full text Add to dashboard Cite

Finding and stopping breaches is now essential in the rapidly evolving world of cybersecurity. This work involves a range of machine learning algorithms for intrusion detection systems (IDS) as well as a detailed examination and use of sophisticated feature engineering techniques. The study enhances intrusion detection's accuracy and efficacy to counteract the ever-changing nature of cyber threats. We handle missing values, encode categorical characteristics, and prepare and balance datasets as part of our strategy. By examining them, one can have a thorough understanding of the benefits and drawbacks of several machine learning models, including Decision Trees, Random Forests, Naive Bayes, and XGBoost, in the context of intrusion detection. Furthermore, we provide a novel approach that combines Recursive Feature Elimination (RFE) with Random Forests to show how well it maintains high classification accuracy while reducing dimensionality. We test the models on a publically available dataset, and we meticulously evaluate the models based on key performance metrics like accuracy, precision, recall, and training time. A comparison analysis of the models makes it clear if they are suitable for real-world application. Our findings indicate that the XGBoost classifier outperforms other models when paired with a well-selected feature set. This demonstrates how well the model detects common intrusions as well as backdoors. In addition, the Random Forest Classifier with RFE modifications exhibits potential as a feature selection technique, offering a balance between accuracy and computational economy. The work adds to the body of knowledge on intrusion detection and provides cybersecurity experts and system developers with valuable information. The strategy being offered offers a reliable and adaptable means of fending off the always changing cyberthreats, thereby bolstering the ongoing effort to fortify digital systems against malicious activity. This work highlights the importance of combining state-of-the-art feature engineering approaches with intricate machine learning algorithms to produce robust and effective cybersecurity defenses, laying a solid foundation for future advancements in intrusion detection.

show abstract

Section: Random Forest Classifiermentioning

confidence: 99%

GuardianAI: Smart Intrusion Detection for Modern Threats

SAIRAJKUMAR

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…This section recalls the works [7][8][9][10][11][12][13][14][15][16][17][18][19][20] from which inspiration was taken for the application of ML models and to make a comparison between the state-of-the-art results and the results achieved by the techniques we presented in the next sections.…”

Section: Related Workmentioning

confidence: 99%

“…In [20], the performances of some supervised (i.e., KNN and SVM) and unsupervised (i.e., isolation forest and k-means) algorithms were evaluated for intrusion detection, using dataset UNSW-NB12.…”

Section: Related Workmentioning

confidence: 99%

Analysis, Design, and Comparison of Machine-Learning Techniques for Networking Intrusion Detection

Dini

Saponara

2021

Designs

View full text Add to dashboard Cite

The use of machine-learning techniques is becoming more and more frequent in solving all those problems where it is difficult to rationally interpret the process of interest. Intrusion detection in networked systems is a problem in which, although it is not fundamental to interpret the measures that one is able to obtain from a process, it is important to obtain an answer from a classification algorithm if the network traffic is characterized by anomalies (and hence, there is a high probability of an intrusion) or not. Due to the increased adoption of SW-defined autonomous systems that are distributed and interconnected, the probability of a cyber attack is increased, as well as its consequence in terms of system reliability, availability, and even safety. In this work, we present the application of different machine-learning models to the problem of anomaly classification in the context of local area network (LAN) traffic analysis. In particular, we present the application of a K-nearest neighbors (KNN) and of an artificial neural network (ANN) to realize an algorithm for intrusion detection systems (IDS). The dataset used in this work is representative of the communication traffic in common LAN networks in military application in particular typical US Air Force LAN. This work presents a training phase of the different models based on a multidimensional-scaling preprocessing procedure, based on different metrics, to provide higher performance and generalization with respect to model prediction capability. The obtained results of KNN and ANN classifiers are compared with respect to a commonly used index of performance for classifiers evaluation.

show abstract

“…Unsupervised and supervised approaches for the identification of outliers were investigated by Laskov et al [2]. A technique for intrusion detection based on outliers was proposed by Portela et al [4].The system makes use of both supervised and unsupervised learning techniques in machine learning. In their empirical investigation, they employ KNN and SVM.…”

Section: Introductionmentioning

confidence: 99%

A hybrid approach for efficient outlier detection using supervised and unsupervised techniques

Jayaramulu,

Venkateswarlu

2024

Preprint

View full text Add to dashboard Cite

Due to data imbalance and dimensionality, it is difficult to achieve optimal performance when detecting outliers in high-dimensional data. Numerous algorithms were developed in try to solve this issue. However, they have their advantages in identifying outliers from such data and are created using either supervised learning technique or unsupervised learning. While unsupervised learning techniques offer mechanisms for discovering and utilising complicated patterns, supervised learning techniques make use of training data. This paper's key premise is that you may "combine two methodologies to create a hybrid and reap the benefits of both worlds." We put forth a cutting-edge machine learning (ML) framework to evaluate this claim, combining supervised and unsupervised techniques for effective outlier detection. Additionally, we suggested an approach called the Multi-Model Approach for Outlier Detection (MMA-OD). The technique improves performance by utilising the advantages of both supervised and unsupervised learning models. Its strength is getting a better feature space. With several benchmark datasets, the suggested approach is assessed. According to the empirical findings, MMA-OD performs better than many other techniques.

show abstract

Evaluation of the performance of supervised and unsupervised Machine learning techniques for intrusion detection

Cited by 9 publications

References 28 publications

GuardianAI: Smart Intrusion Detection for Modern Threats

GuardianAI: Smart Intrusion Detection for Modern Threats

Analysis, Design, and Comparison of Machine-Learning Techniques for Networking Intrusion Detection

A hybrid approach for efficient outlier detection using supervised and unsupervised techniques

Contact Info

Product

Resources

About