信息网络安全 ›› 2017, Vol. 17 ›› Issue (6): 62-67.doi: 10.3969/j.issn.1671-1122.2017.06.010

• 技术研究 • 上一篇    下一篇

基于改进Logistic回归算法的抗Web DDoS攻击模型的设计与实现

张雪博1, 刘敬浩1, 付晓梅2   

  1. 1. 天津大学电气自动化与信息工程学院,天津300072;
    2. 天津大学海洋科学与技术学院, 天津300072
  • 收稿日期:2017-05-05 出版日期:2017-06-20
  • 通讯作者: 张雪博 zxb_tju@163.com
  • 作者简介:张雪博(1991-),男,河南,硕士研究生,主要研究方向为网络安全;刘敬浩(1963-),男,天津,副教授,硕士,主要研究方向为网络安全、网络虚拟环境、无线网络通信;付晓梅(1968-),女,重庆,副教授,博士,主要研究方向为无线通信、海洋通信、信息安全。
  • 基金资助:
    国家自然科学基金[61571323]

Design and Implementation of Anti Web DDoS Attack Model Based on Improved Logistic Regression Algorithm

ZHANG Xuebo1, LIU Jinghao1, FU Xiaomei2   

  1. 1. School of Electrical Automation and Information Engineering, Tianjin University, Tianjin 300072, China;
    2.School of Marine Science and Technology, Tianjin University, Tianjin 300072, China
  • Received:2017-05-05 Online:2017-06-20

摘要: Web DDoS攻击已经成为黑客常用的攻击手段之一。为了有效地提高Web DDoS攻击的检测速度和检测率,文章将量子粒子群优化方法与Logistic回归模型相结合,提出了一种轻量级检测新算法。该算法通过自适应的量子粒子群优化方法取代Newton法对Logistic回归系数进行求解,提高了回归系数的求解效率和精度。为了验证本算法的有效性,实验采用WorldCup98公开数据集对本算法与现有的改进Logistic回归算法的性能进行对比分析。实验结果表明,在Web DDoS攻击的检测方面,相比现有的改进Logistic回归算法,文章提出的算法能够获得更高的检测率以及更低的误检率,同时算法的时间复杂度与检测样本数量之间为线性关系。

关键词: Web DDoS攻击检测, Logistic回归, 量子粒子群优化算法, 牛顿法

Abstract: Web DDoS attack has become one of the common ways for hackers to attack. In order to improve the detection speed and accuracy of Web DDoS attack effectively, this paper proposes a light weight and novel detection algorithm combined quantum particle swarm optimization method with Logistic regression model. This algorithm replaces Newton method with adaptive swarm optimization method to solve Logistic regression coefficient, improving the efficiency and accuracy of solving the regression coefficient. In order to verify the availability of the proposed algorithm, the WorldCup98 open dataset was used in our study to compare the performance of our algorithm with the existing improved Logistic regression algorithms.The experimental results show that compared with the existing improved Logistic regression algorithm, the proposed algorithm has higher detection rate and smaller detection error rate in terms of detecting Web DDoS attacks. Meanwhile,there is a linear relationship between the time complexity of the proposed algorithm and the number of detection sample.

Key words: Web DDoS attack detection, Logistic regression, quantum particle swarm optimization algorithm, Newton method

中图分类号: