信息网络安全 ›› 2018, Vol. 18 ›› Issue (5): 1-11.doi: 10.3969/j.issn.1671-1122.2018.05.001

• •    下一篇

基于机器学习的入侵检测方法对比研究

和湘1, 刘晟1(), 姜吉国2   

  1. 1.国防科技大学信息通信学院,湖北武汉 430000
    2. 山东省公安厅,山东济南 250001
  • 收稿日期:2018-03-10 出版日期:2018-05-15 发布日期:2020-05-11
  • 作者简介:

    作者简介:和湘(1976—),女,湖南,副教授,硕士,主要研究方向为无线网络安全、机器学习等;刘晟(1996—),男,福建,本科,主要研究方向为机器学习、深度学习、系统架构;姜吉国(1983—),男,山东,本科,主要研究方向为电子数据勘查取证。

  • 基金资助:
    国家自然科学基金[61601490]

Comparative Study of Intrusion Detection Methods Based on Machine Learning

Xiang HE1, Sheng LIU1(), Jiguo JIANG2   

  1. 1. College of Information and Communication, National University of Defense Technology, Wuhan Hubei 430000, China
    2. Public Security Department of Shandong Province, Jinan Shandong 250001, China
  • Received:2018-03-10 Online:2018-05-15 Published:2020-05-11

摘要:

随着网络安全形势日趋严峻,入侵检测技术已经成为保障网络安全的一种重要手段。因此把机器学习的理论和方法引入入侵检测已成为一种共识,并且近些年来在这一研究领域取得了不错的进展。文章对比分析了不同机器学习方法在入侵检测上的应用。首先,介绍机器学习的一般化过程,对典型机器学习方法的理论进行对比分析。然后,对不同机器学习方法做仿真研究,观察性能变化。最后,在仿真的基础上对不同模型进行横向比较。文章在仿真实验的基础上得出了较为可靠的结论,对找出具有性能优势的机器学习方法具有重要意义。

关键词: 入侵检测, 机器学习, 决策树, 支持向量机, 神经网络

Abstract:

With the network security situation becomes more and more severe, intrusion detection technology has already become an important means to ensure network security. Therefore, it has become a consensus to introduce the theory and method of machine learning to intrusion detection. In recent years, considerable progress has been made in this field. The article analyzes the application of different machine learning methods in intrusion detection. First of all, The article introduces the general process of machine learning and compares and analyzes the theories of typical machine learning methods. Then the article uses different machine learning methods for simulation study to observe the performance changes. Finally, the article carries out the horizontal comparison of different models on the basis of simulation. Based on the simulation experiments, the article draws a more reliable conclusion, which is of great significances to search for a machine learning algorithm which has better performances.

Key words: intrusion detection, machine learning, decision tree, support vector machine, neural network

中图分类号: