In response to the problems of traditional intrusion detection methods for high-dimensional traffic data, such as data imbalance, low feature extraction efficiency, and difficulty in convergence during training, an intrusion detection method based on deep convolutional neural networks combined with data balancing is proposed. Firstly, the CICIDS-2018 network traffic dataset is transformed into grayscale images with integer values, which are then input into a conditional generative adversarial network. Secondly, the trained generator is used to produce attack data for the minority class, which is added to the original dataset to balance it. Finally, the performance of intrusion detection is enhanced by using deep convolutional neural network. Experimental results show that the accuracy rate in multi-classification task of this method is 96.58%, which is superior to that of traditional detection methods, with the classification accuracy rates for the two least frequent attack traffic data Botnet and SQL increased by 5.83% and 32.18%, respectively, compared to those before balancing.
CICIDS-2018(Canadian Institute for Cybersec-urity Intrusion Detection Systems Evaluation Dataset 2018)是一个广泛应用于网络安全领域的公开数据集,旨在帮助研究人员和工业界专家开发和评估入侵检测系统。正常类别(Benign)以及异常攻击(其他)标签类型统计如表1所示。
AKHTAR MALI, QADRIS M O, SIDDIQUIM A, et al. Robust genetic machine learning ensemble model for intrusion detection in network traffic[J]. Scientific Reports, 2023,13:No.17227.
[2]
ABDULHAMMEDR, FAEZIPOURM, ABUZNEIDA, et al. Deep and machine learning approaches for anomaly-based intrusion detection of imbalanced network traffic[J]. IEEE Sensors Letters, 2019,3(1):No.7101404.
SHARMAH S, SARKARA, SINGHM M. An efficient deep learning-based solution for network intrusion detection in wireless sensor network[J]. International Journal of System Assurance Engineering and Management, 2023,14(6):2423-2446.
DENGM L, SUNC C, KANY P, et al. Network intrusion detection based on deep belief network broad equalization learning system[J]. Electronics, 2024,13(15):No.3014.
[16]
JOSEJ, JOSED V. Deep learning algorithms for intrusion detection systems in Internet of Things using CIC-IDS 2017 dataset[J]. International Journal of Electrical and Computer Engineering, 2023,13(1):No.1134.
[17]
CHENY, LINQ Z, WEIW H, et al. Intrusion detection using multi-objective evolutionary convolutional neural network for Internet of Things in Fog computing[J]. Knowledge-Based Systems, 2022,244:No.108505.