Securing 6G-enabled IoT/IoV networks by machine learning and data fusion

Abstract

The rapid growth of Internet of Things (IoT) and Internet of Vehicles (IoV) are rapidly moving to the 6G networks, which leads to dramatically raised security issues. Using machine learning, including deep learning, to find out malicious network traffic is one of practical ways. Though much work has been done in this direction, we found little investigating the effect of using fused network conversation datasets to train and test models. Thus, this work proposes to check conversation dataset characteristics and find suitable ones to fuse into one dataset in order to improve the capability of malicious traffic and malware detection performance. The experiments using real data show that conditioned combination of datasets can be used to enhance algorithm performance and improve detection results. For this reason, it is recommended to profile datasets and conduct conditional fusion of network conversation datasets before using machine learning or deep learning. As the characterization is done using general statistical calculation, it is promising to be used for other domains too.

Artificial intelligence enabled Luong Attention and Hosmer Lemeshow Regression Window‐based attack detection in 6G

“…In Lopez-Martin et al, 15 conversation dataset features were analyzed and identified pertinent features with the purpose of enhancing the potentiality of malicious traffic and also improving the malware detection performance significantly.…”

“…From this result it is inferred that the attack detection time using the LA-HLRW method was found to be comparatively lesser than. 1,2,9 The dynamic attention mechanism employed in the proposed LA-HLRW method integrates the attention function for activating the hidden layer with the purpose of improving the overall LSTM network performance. This in turn not only results in reducing dimensionality reduced network traffic feature matrix but also results in minimum convergence.…”

“…However, while observing the simulation the attack detection accuracy using the three methods was observed to be 90.35%, 85.45%, 81.45%, and 83.65% respectively using LA-HLRW. 1,2,9 The reason for the improvement was first to strengthen the information extraction (i.e., network traffic feature extraction), attention mechanism was applied to the hidden layer activation function. This in turn assisted in extracting the most significant feature that in turn reduced the dimensionality also.…”

“…From this result, the precision rate using was observed to be comparatively better than. 1,2,9 The reason was due to the application of the Hosmer Lemeshow Logistic Regression Window-based Attack Detection algorithm for attack detection. In this algorithm by combining the Hosmer Lemeshow Logistic Regression function to the sliding window adjusted the likelihood threshold for predicting abnormalities in an efficient manner.…”

“…Along with the performance metric precision, recall is said to be another parameter of significance for analyzing the attack detection efficiency. The recall rate using LA-HLRW and existing methods of machine learning and data fusion, 1 Weight Based Ensemble machine Learning Algorithm (WBELA), 2 and Attention Mechanism and Auto Encoder for Intrusion Detection (AMAEID) 9 are given in Table 6. Finally, Figure 9, given above shows the comparative results of recall rate using three methods, LA-HLRW, Machine learning and data fusion, 1 WBELA, 2 and AMAEID 9 respectively.…”