Real-Time Intrusion Detection in Wireless Network: A Deep Learning-Based Intelligent Mechanism

Abstract

With the development of the wireless network techniques, the number of cyber-attack increases significantly, which has seriously threat the security of Wireless Local Area Network (WLAN). The traditional intrusion detection technology is a prevalent area of study for numerous years, but it may not have a good detection performance in a real-time way. Therefore, it is urgent to design a detection mechanism to detect the attacks timely. In this paper, we exploit a CDBN (Conditional Deep Belief Network)-based intrusion detection mechanism to recognize the attack features and perform the wireless network intrusion detection in real time. To avoid the impact of the imbalanced dataset and the data redundancy on the detection accuracy, a windowbased instance selection algorithm-SamSelect‖ is adopted to undersample the majority class data samples, and a Stacked Contractive Auto-Encoder (SCAE) algorithm is proposed to reduce the dimension of the data samples. By doing so, our proposed mechanism can effectively detect the potential attack and achieve high accuracy. The experiment results show that CDBN can be effectively combined with-SamSelect‖ and SCAE, and the proposed mechanism has a high detection speed and accuracy, with the average detection time 1.14 ms and the detection accuracy 0.974.

Wireless Local Area Networks Threat Detection Using 1D-CNN

“…This could be attributed to the imbalance in the testing sets used. Only SamSelect+SCAE+CDBN [ 13 ] and Support Vector Machines [ 17 ] addressed this issue by balancing their testing datasets. Other approaches that outperformed our models in terms of accuracy did not balance their testing datasets, which led to higher accuracy scores, as one class made up a majority of the data set.…”

“…Thus, as shown in the previous section, it consumes minimal amounts of energy. Upon analyzing Table 12 , it can be seen that our two models, both binary and multi-class, perform slightly worse than the SamSelect+SCAE+CDBN deep learning model from [ 13 ] and the Support Vector Machines from [ 17 ] when it comes to F1-score. Despite that, it is worth mentioning that our models are surely less complex than these solutions which is definitely an advantage considering the slightly worse F1-score.…”

“…Our multi-class model achieves a better F1-score, indicating that our model is more balanced and reliable in predicting all classes. Considering AUC, only two papers provided this metric, nevertheless proposed approach achieves the best results with the multi-class model on top with a 0.99 AUC score and a binary model with a 0.972 score while Random Forest [ 12 ] and SamSelect+SCAE+CDBN [ 13 ] scored 0.704 and 0.978 respectively. This means that the multi-class model is the best in distinguishing different classes.…”

“…The authors of [ 13 ] propose a deep learning model for classification. They use a window-based instance selection algorithm called “SamSelect” to balance data, and a stacked Contractive Auto-Encoder to reduce dimensionality.…”

An Ensemble Intrusion Detection Method for Train Ethernet Consist Network Based on CNN and RNN

“…They developed a new approach called the weighted one-againstrest SVM based on the model selection method. Yang et al [15] exploited a conditional deep belief network-based IDS to perform the wireless network intrusion detection in real time. Their experiment results show that their method has a high detection speed and accuracy with the average detection time 1.14 ms and the detection accuracy 0.974.…”