《计算机应用》唯一官方网站 ›› 2023, Vol. 43 ›› Issue (8): 2471-2476.DOI: 10.11772/j.issn.1001-9081.2022071030

• 网络空间安全 • 上一篇    

基于改进Inception-ResNet的加密流量分类方法

郭祥, 姜文刚(), 王宇航   

  1. 江苏科技大学 自动化学院,江苏 镇江 212100
  • 收稿日期:2022-07-14 修回日期:2022-11-17 接受日期:2022-11-21 发布日期:2023-01-15 出版日期:2023-08-10
  • 通讯作者: 姜文刚
  • 作者简介:郭祥(1997—),男,安徽六安人,硕士研究生,主要研究方向:多媒体与信息安全、深度学习
    王宇航(1996—),女,吉林长春人,硕士,主要研究方向:多媒体与信息安全、深度学习。
  • 基金资助:
    国家自然科学基金资助项目(61702235);江苏省研究生创新计划项目(KYCX21_3482)

Encrypted traffic classification method based on improved Inception-ResNet

Xiang GUO, Wengang JIANG(), Yuhang WANG   

  1. College of Automation,Jiangsu University of Science and Technology,Zhenjiang Jiangsu 212100,China
  • Received:2022-07-14 Revised:2022-11-17 Accepted:2022-11-21 Online:2023-01-15 Published:2023-08-10
  • Contact: Wengang JIANG
  • About author:GUO Xiang, born in 1997, M. S. candidate. His research interests include multimedia and information security, deep learning.
    WANG Yuhang, born in 1996, M. S. Her research interests include multimedia and information security, deep learning.
  • Supported by:
    National Natural Science Foundation of China(61702235);Graduate Innovation Program of Jiangsu Province(KYCX21_3482)

摘要:

基于深度学习的加密流量分类方法中的分类模型大多是深层直筒型结构,存在梯度消失的问题,且网络层数的增加会使模型结构和计算的复杂度显著上升。为此,提出了一种基于改进Inception-ResNet的加密流量分类方法。该方法通过改进Inception模块,并将该模块作为残差块以残差结构连接的方式嵌入卷积神经网络来构建分类模型;此外,改进分类模型的损失函数,并使用VPN-nonVPN数据集来验证所提方法的有效性。实验结果表明,所提方法在2种场景的分类实验中的精确率、召回率、F1值分别达到了94.21%、92.53%和93.31%以上。在与其他方法的对比实验中,以分类难度最大的12分类实验为例,所提方法比C4.5决策树算法和1D-CNN(1 Dimensional-Convolutional Neural Network)在精确率上分别高出13.91和9.50个百分点,在召回率上分别高出14.87和1.59个百分点。与CAE (Convolutional Auto Encoding)和SAE (Stacked Auto Encoder)等方法相比,所提方法虽然在各项指标上没有明显提升,但在单次训练时长上却有明显缩短,充分表明了所提方法的先进性。

关键词: 深度学习, 批量归一化层, 残差结构, 不平衡数据集, 损失函数

Abstract:

Most classification models in deep learning-based encrypted traffic classification methods have deep and straight structure with the problem of vanishing gradient, and the increase of the number of network layers leads to significant increase of model structure and computational complexity. Based on these, an encrypted traffic classification method based on improved Inception-ResNet was proposed. In the method, the classification model was constructed by improving the Inception module and embedding it into the convolutional neural network as a residual block in a residual structural connection way. In addition, the loss function of the classification model was improved, and the effectiveness of the proposed method was verified by using VPN-nonVPN dataset. Experimental results show that the proposed method achieves the precision, recall, and F1 score of more than 94.21%, 92.53%, and 93.31%, respectively, in the classification experiments of two senerios. In the comparison experiments with other methods, taking the 12-class classification experiment, which is the most difficult one, as an example, the proposed method is higher than C4.5 decision tree algorithm and 1D-CNN (1 Dimensional-Convolutional Neural Network) by 13.91 and 9.50 percentage points higher in precision and by 14.87 and 1.59 percentage points in recall. Compared with the algorithms such as CAE (Convolutional Auto Encoding) and SAE (Stacked Auto Encoder), the proposed method not has obvious improvement on the indicators, but has significant shorter single training time, fully demonstrating that the proposed method is a state-of-the-art method.

Key words: deep learning, batch normalization layer, residual structure, imbalanced dataset, loss function

中图分类号: