信息网络安全 ›› 2018, Vol. 18 ›› Issue (8): 73-78.doi: 10.3969/j.issn.1671-1122.2018.08.010

• • 上一篇    下一篇

非负矩阵分解算法优化及其在入侵检测中的应用

张戈琳1, 李勇1,2()   

  1. 1. 北京交通大学电子信息工程学院,北京 100044
    2. 广西密码学与信息安全重点实验室,广西桂林 541004
  • 收稿日期:2017-12-10 出版日期:2018-08-20 发布日期:2020-05-11
  • 作者简介:

    作者简介:张戈琳(1993—),女,内蒙古,硕士研究生,主要研究方向为网络入侵;李勇(1973—),男,山东,副教授,博士,主要研究方向为应用密码学、云计算与大数据安全等。

  • 基金资助:
    国家自然科学基金面上项目[61472032];广西密码学与信息安全重点实验室研究课题[GCIS201609]

Non-negative Matrix Factorization Optimization and Its Application in Network Intrusion Detection

Gelin ZHANG1, Yong LI1,2()   

  1. 1. School of Electronic and Information Engineering, Beijing Jiaotong University, Beijing 100044, China
    2. Guangxi Key Laboratory of Cryptography and Information Security, Guilin Guangxi 541004, China
  • Received:2017-12-10 Online:2018-08-20 Published:2020-05-11

摘要:

文章将NMF算法与主成分分析算法相结合,对NMF算法的初始化问题进行优化,并应用于入侵检测中。针对如何确定保留基的个数K和NMF矩阵初始化的问题,文章研究了改进后的NMF算法在网络入侵领域的应用。为实现定性分析,文章将KDD数据集记录从高维空间降至低维空间,然后在低维空间中展示数据特征。为实现定量分析,文章将数据通过SVM进行分类处理,生成检测报告,证明了改进后的NMF算法在检测率和效率上比原有算法有所提高。

关键词: 网络安全, 入侵检测, 非负矩阵分解算法, SVM

Abstract:

Due to the non-negative matrix factorization (NMF) can effectively reduce the high-dimensional data to the low-dimension by decomposition, the initialization problem of non-negative matrix factorization algorithm was optimized by combining principal component analysis algorithm, and then it was applied to intrusion detection. For the problem of how to determine the number of reserved bases K and NMF matrix initialization, the application of the improved NMF algorithm in the field of network intrusion is studied. In order to achieve qualitative analysis, we reduce KDD dataset records from high-dimensional space to low-dimensional space, and then display data features in low-dimensional space. To achieve quantitative analysis, the data is classified by SVM and processed into test reports to verify that the optimized NMF algorithm is better than the original algorithm in detection rate and efficiency.

Key words: network security, intrusion detection, non-negative matrix factorization, SVM

中图分类号: