Abstract
Network Intrusion Detection System (NIDS) detects normal and malicious behavior by analyzing network traffic, this analysis has the potential to detect novel attacks especially in IoT environments. Deep Learning (DL)has proven its outperformance compared to machine learning algorithms in solving the complex problems of the real-world like NIDS. Although, this approach needs more computational resources and consumes a long time. Feature selection plays a significant role in choosing the best features only that describe the target concept optimally during a classification process. However, when handling a large number of features the selecting such relevant features becomes a difficult task. Therefore, this paper proposes Enhanced BPSO using Binary Particle Swarm Optimization (BPSO) and correlation–based (CFS) classical statistical feature selection approach to solve the problem on BPSO feature selection. The selected feature subset has evaluated on Deep Neural Networks (DNN) classifiers and the new flow-based CSE-CIC-IDS2018 dataset. Experimental results have shown a high accuracy of 95% based on processing time, detection rate, and false alarm rate compared with other benchmark classifiers.
Highlights
The exponentially growing number of security breaches, cyberattacks on Internet of things IOT highly required reliable security solutions
They proposed a new malware prediction model that could detect the coming future malware by the implementing a deep learning method of Mal Generative Adversarial Network (Mal-generative adversarial networks (GANs)) [4]. showed that the LSTM classifier outperform over previously published results of other static classifiers on KDD Cup '99 dataset challenge for long time which prove the benefit of LSTM networks to intrusion detection, because the ability of LSTM to learn from look back in time and link connection records consecutively [5].The recurrent neural networks (RNNs), Stacked RNN, and convolutional neural networks (CNNs) are supervised deep learning techniques applied to classify common five attack types using Keras .This technique used packet header information without need any user payload compared its results with Snort IDS .The results showed that this technique gave superior results compared Snort [6]
Used the genetic operatorson the Particle Swarm Optimization algorithm (PSO) to search the global solution for optimization which used to construct the network structure for intrusion detection on NSL-KDD[14].The researcher aimed to improving the performance of Network Intrusion Detection System (NIDS) on UNSW- N15 dataset by proposed four feature selection models based on the particle swarm optimization (PSO), firefly optimization (FFA),genetic algorithm (GA)and grey wolf optimizer (GWO).The derived features from this model are evaluated on the J48 machine learning (ML) and support vector machine(SVM) classifiers[15].A double PSO-based algorithm proposed to select subset of features and hyper parameters both in the same work
Summary
The exponentially growing number of security breaches, cyberattacks on Internet of things IOT highly required reliable security solutions. Traditional machine learning (ML) approaches have been supplied for cyber security such as Bayesian Belief Networks (BBN), Random forest, Support Vector Machines (SVM) and others, but the generation of large scale data in IoT required a deep learning based approach which performs better with large data sizes and can learn representation of feature from raw data so it is adaptable to different attack scenarios [3] They proposed a new malware prediction model that could detect the coming future malware by the implementing a deep learning method of Mal Generative Adversarial Network (Mal-GAN) [4]. Improving the NIDS by optimized deep learning models with pre-processing phase employing a Binary PSO algorithm This approach optimized detection rate (DR) of deep learning models while reducing false alarm rate(FAR)compared with corresponding values of deep learning models without preprocessing phase. Evaluating this approach by using new CSE-CIC-IDS2018 real datasets for classification tasks. Deep learning algorithms are often limited by weak points of data and parameters
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
