为了提高舰船的网络安全，降低网络攻击的影响，提出深度学习算法的舰船网络安全状态识别方法。采集舰船网络流量数据，使用随机森林算法计算舰船网络流量数据特征基尼系数，以此为标准比较、排序每一个特征在随机森林中的贡献平均值，得出不同特征的重要程度，将重要程度高的特征输入BiLSTM神经网络之中，利用神经网络的自我学习完成对舰船的网络安全状态的识别，并且增加注意力机制进一步提高学习效率和分类准确率。实验数据表明，在利用随机森林进行特征提取时，特征数量选择为75个时，可兼顾特征充分参与分支、计算效率与识别准确度；网络受到攻击后流量有明显的改变，根据网络流量改变情况，使用该方法能够判断出船网络遭受的攻击方式。
In order to improve the network security of ships and reduce the impact of network attacks on ship networks, a deep learning algorithm for ship network security state recognition is proposed. Collect ship network traffic data, use the random forest algorithm to calculate the Gini coefficient of ship network traffic data features, and use this as a standard to compare and sort the average contribution of each feature in the random forest to determine the importance of different features. Input the highly important features into the BiLSTM neural network, and use the self-learning of the neural network to complete the recognition of the network security status of ships, and increasing attention mechanisms further improves learning efficiency and classification accuracy. Experimental data shows that when using random forests for feature extraction, when the number of features is chosen to be 75, it can balance the full participation of features in branching, computational efficiency, and recognition accuracy. After the ship network is attacked, there is a significant change in network traffic. Based on the changes in network traffic, this method can be used to determine the attack mode suffered by the ship network.
2023,45(21): 193-196 收稿日期：2023-4-10
DOI：10.3404/j.issn.1672-7649.2023.21.037
分类号：TP309
作者简介：杨春霞(1980-)，女，硕士，副教授，主要从事数据通信与网络技术研究
参考文献：
[1] 熊中浩, 张伟, 杨国玉. 基于DBN的网络安全态势评估和态势预测建模研究[J]. 电子技术应用, 2021, 47(5): 35-39+44.XIONG Zhonghao, ZHANG Wei, YANG Guoyu. Research on network security situation assessment and situation prediction modeling based on DBN [J]. Application of Electronic Technique, 2021, 47(5): 35-39+44.
[2] 罗予东, 陆璐. 基于人工神经网络和遗传算法的网络攻击检测[J]. 计算机工程与设计, 2021, 42(9): 2446-2454.LUO Yudong, LU Lu. Network attack detection based on artificial neural networks and genetic algorithm [J]. Computer Engineering and Design , 2021, 42(9): 2446-2454.
[3] 杨宏宇, 张旭高. 一种网络安全态势自适应预测模型[J]. 西安电子科技大学学报, 2020, 47(3): 14-22.YANG Hongyu, ZHANG Xugao. Network security situation adaptive prediction model[J]. Journal of Xidian University(Natural Science), 2020, 47(3): 14-22.
[4] 王婷, 王娜, 崔运鹏, 等. 基于半监督学习的无线网络攻击行为检测优化方法[J]. 计算机研究与发展, 2020, 57(4): 791-802.WANG Ting, WANG Na, CUI Yunpeng, et al. The Optimization Method of Wireless Network Attacks Detection Based on Semi-Supervised Learning[J]. Journal of Computer Research and Development, 2020, 57(4): 791-802.
