Employing Intelligent Algorithms and Deep Learning for Network Intrusion Detection

Research areas:

• Volume 11,Issue 3, Nov. 2022

Year:

2022

Type of Publication:

Article

Keywords:

Convolution Neural Network, Principal Component Analysis, Singular Value Decomposition, Network Intrusion Detection Convolution Neural Network, UNSW-NB15

Authors:

Inbithaq A. Shakir; Hazem M. El-Bakry; Ahmed A. Al-fetouh Saleh

Journal:

IJAIM

Volume:

11

Number:

3

Pages:

34-47

Month:

November

ISSN:

2320-5121

Abstract:

Any corporation may benefit from having electronic information, and even an individual may value their data to the point that they cannot afford to lose it. In today’s digital age, information security has grown increasingly crucial, necessitating viable defenses against ever-evolving threats. As a result, cybersecurity and risk management are critical for occupations that rely on data or information. Deep Learning (DL) methods have been widely used in information security in recent years. Because of unexpected behavior and undiscovered weaknesses, traditional rule-based security systems are vulnerable to sophisticated assaults. In this study, a novel deep Convolution Neural Network (CNN) model for intrusion detection was created. The proposed method uses two feature reduction approaches, Principal Component Analysis (PCA) and Singular Value Decomposition, together with a number of preprocessing processes (SVD). With a 100% accuracy rate, the investigations employing the UNSW-NB15 dataset produced encouraging findings.

Full text: IJAIM_662_FINAL.pdf