A New Intrusion Detection System for the Internet of Things via Deep Convolutional Neural Network and Feature Engineering

Ullah, Safi; Ahmad, Jawad; Khan, Muazzam A.; Alkhammash, Eman H.; Hadjouni, Myriam; Ghadi, Yazeed Yasin; Saeed, Faisal; Pitropakis, Nikolaos

doi:10.3390/s22103607

Cited by 34 publications

(18 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this method, multiple traditional classifiers are combined as basic learners, and the prediction of traditional classifiers is voted to obtain the final prediction. Ullah et al [24] proposed an intrusion detection system based on deep convolution neural network. The model consists of two convolutional layers and three fully connected dense layers, which can improve performance and reduce computing power.…”

Section: Related Workmentioning

confidence: 99%

“…LSTM and RNN models are mainly used to solve the problem of gradient disappearance in the process of back propagation. They are good at short-term memory, but poor in long-term memory and in the face of apt attacks and bot attacks [24], so it is difficult to train efficient IOT attack prevention models. As shown in Fig.…”

Section: Cloud-based Loss Function Computational Designmentioning

confidence: 99%

See 1 more Smart Citation

Internet of Things Intrusion Detection System Based on Convolutional Neural Network

Yin¹,

Shi²,

Dong³

et al. 2023

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

In recent years, the Internet of Things (IoT) technology has developed by leaps and bounds. However, the large and heterogeneous network structure of IoT brings high management costs. In particular, the low cost of IoT devices exposes them to more serious security concerns. First, a convolutional neural network intrusion detection system for IoT devices is proposed. After cleaning and preprocessing the NSL-KDD dataset, this paper uses feature engineering methods to select appropriate features. Then, based on the combination of DCNN and machine learning, this paper designs a cloud-based loss function, which adopts a regularization method to prevent overfitting. The model consists of one input layer, two convolutional layers, two pooling layers and three fully connected layers and one output layer. Finally, a framework that can fully consider the user's privacy protection is proposed. The framework can only exchange model parameters or intermediate results without exchanging local individuals or sample data. This paper further builds a global model based on virtual fusion data, so as to achieve a balance between data privacy protection and data sharing computing. The performance indicators such as accuracy, precision, recall, F1 score, and AUC of the model are verified by simulation. The results show that the model is helpful in solving the problem that the IoT intrusion detection system cannot achieve high precision and low cost at the same time.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Cloud-based Loss Function Computational Designmentioning

confidence: 99%

Internet of Things Intrusion Detection System Based on Convolutional Neural Network

Yin¹,

Shi²,

Dong³

et al. 2023

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

show abstract

“…A deep neural network (DNN) feature extractor that collects beneficial properties from a dynamic distributed system, a kmeans clustering that groups the collected information, and a PPO operator that automate the IDS via retraining and command ID-HyConSys' intrusion detection module. A deep-convolutional neural network (DCNN)-based IDS is suggested in [11]. Three completely associated dense tiers and two pooling layers make up such model.…”

Section: Related Workmentioning

confidence: 99%

“…is the existing answer for the iterations, and is the -th searching solutions. In order to balance the search strategy, a quality function known as QF is determined using Equation (11). (11) The variables and depict the motions of the Aquila's prey, with 's value dropping 2 to 0.…”

Section: Feature Selection Using Aquila Optimizer Algorithmmentioning

confidence: 99%

Network Intrusion Detection Method Using Stacked BILSTM Elastic Regression Classifier with Aquila Optimizer Algorithm for Internet of Things (IoT)

Gangula¹,

Vutukuru²,

M.³

2023

IJRITCC

View full text Add to dashboard Cite

Globally, over the past ten years, computer networks and Internet of Things (IoT) networks have grown significantly due to the increasing amount of data that has been collected, ranging from zettabytes to petabytes. As a result, as the network has expanded, security problems have also emerged. The large data sets involved in these types of attacks can make detection difficult. The developing networks are being used for a multitude of sophisticated purposes, such as smart homes, cities, grids, gadgets, and objects, as well as e-commerce, e-banking, and e-government. As a result of the development of numerous intrusion detection systems (IDS), computer networks are now protected from security and privacy threats. Data confidentiality, integrity, and availability will suffer if IDS prevention efforts fail. Complex attacks can't be handled by traditional methods. There has been a growing interest in advanced deep learning techniques for detecting intrusions and identifying abnormal behavior in networks. This research aims to propose a novel network namely stacked BiLSTM elastic regression classifier (Stack_BiLSTM-ERC) with Aquila optimizer algorithm for feature selection. This optimization method computes use of a cutting-edge transition function that enables it to be transformed into a binary form of the Aquila optimizer. A better solution could be secured once number of possible solutions are found from diverse regions of the search space utilizing the Aquila optimizer method. NSL-KDD and UNSW-NB15 are two datasets that enable learning characteristics from the raw data in order to detect harmful prerequisites characteristics and effective framework patterns. The proposed Stack_BiLSTM-ERC achieves 98.l3% of accuracy, 95.1% of precision, 94.3% of recall and 95.4 of F1-score for NSL-KDD dataset. Moreover, 98.6% of accuracy, 97.2% of precision, 98.5 of recall and 97.5% of F1-score.

show abstract

“…As a result of the classification, an accuracy rate of 87.6% was obtained. In the Deep Convolutional Neural Network (DCNN) based IDS proposed by Ullah et al [20], 62 features with an Information Gain (IG) value greater than 0.001 were selected by IG method among 83 features in the IoTID20 dataset. In the intrusion detection method conducted with DCNN, 99.84% accuracy was obtained for binary classification and 98.12% accuracy was achieved for multi-class classification.…”

Section: Literature Reviewmentioning

confidence: 99%

A Hybrid PCA-MAO Based LSTM Model for Intrusion Detection in IoT Environments

Karamollaoglu

Yücedağ

Doğru

2022

Preprint

View full text Add to dashboard Cite

With the rapid advances in Internet of Things (IoT) technologies, the number of smart objects connected to IoT networks is increasing day by day. Parallel to this exponential growth, attacks against IoT networks are also increasing rapidly. Various Intrusion Detection Systems (IDS) have been proposed by researchers to improve accuracy in detecting attacks with different behaviors and reduce intrusion detection time. This work presents a novel IDS based on the combination of the Principal Component Analysis and Mayfly Optimization methods (PCA-MAO) for dimensionality reduction, the Borderline Synthetic Minority Oversampling Technique (BSMOTE) for data balancing, and the Long Short-Term Memory (LSTM) method for classification. A new dataset was created by combining IoTID20, CIC-ToN-IoT and USB-IDS-1 datasets to be used in the performance test of the proposed model. Thus, the performance evaluation of the proposed model was performed for more attack types with different behaviors. As a result of classification using the proposed hybrid PCA-MAO based LSTM model, an accuracy of 99.51% was obtained. It has been observed that the proposed IDS provides superior intrusion detection performance for high-dimensional, complicated, and imbalanced data compared to classical machine learning (ML) methods.

show abstract

A New Intrusion Detection System for the Internet of Things via Deep Convolutional Neural Network and Feature Engineering

Cited by 34 publications

References 32 publications

Internet of Things Intrusion Detection System Based on Convolutional Neural Network

Internet of Things Intrusion Detection System Based on Convolutional Neural Network

Network Intrusion Detection Method Using Stacked BILSTM Elastic Regression Classifier with Aquila Optimizer Algorithm for Internet of Things (IoT)

A Hybrid PCA-MAO Based LSTM Model for Intrusion Detection in IoT Environments

Contact Info

Product

Resources

About