Work

Efficient Feature Engineering-Based Anomaly Detection for Network Security

Year: 2024
Type: article
Abstract: With the rapid advancement of internet technology, network-based attacks have become increasingly prevalent, posing significant challenges to ensuring the security of network infrastructures. In response, feature selection and feature reduction have emerged as essential techniques for dealing with the large volumes of data inherent in network security applications. However, traditional feature selection methods may not always suffice when all attributes are crucial for anomaly detection. To address this, we propose a feature engineering-based approach that combines feature selection and feature extraction to effectively reduce dimensionality while preserving relevant attributes. Specifically, we introduce a Stochastic-based Feature Engineering (S_FE) algorithm tailored for both manual packet capture and real-time payload datasets. In manual packet capture datasets, our algorithm extracts Trust Value, Byte Frequency Analysis (BFA), Byte Entropy (BE), Payload Length (PL), and Stream Index features, while for real-time payload datasets, it focuses on Trust Value, direction, and Hash Value features. We compare the performance of our S_FE algorithm against widely used Feature Engineering (FE) algorithms using key metrics such as accuracy, precision, recall, and F1-score. Experimental results demonstrate the superior performance of our proposed algorithm, highlighting its efficacy in network anomaly detection. This research contributes to the development of efficient techniques for enhancing network security in the face of evolving cyber threats.
Author: Mayukha S
Language: English
Open Access status: diamond
OpenAlex ID: W7217573092
FWCI: 0.321
Cites: 30
Cited by: 3
Related to: 10