通信学报 ›› 2016, Vol. 37 ›› Issue (1): 42-48.doi: 10.11959/j.issn.1000-436x.2016006

• 学术论文 • 上一篇    下一篇

基于改进FastICA算法的入侵检测样本数据优化方法

杜晔,张亚丹,黎妹红,张大伟   

  1. 北京交通大学计算机与信息技术学院,北京 100044
  • 出版日期:2016-01-25 发布日期:2016-01-27
  • 基金资助:
    国家自然科学基金资助项目

Improved FastICA algorithm for data optimization processing in intrusion detection

Ye DU,dan ZHANGYa,hong LIMei,wei ZHANGDa   

  1. School of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044,China
  • Online:2016-01-25 Published:2016-01-27
  • Supported by:
    Beijing Higher Education Young Elite Teacher Project

摘要:

为更好实现对入侵检测样本数据的优化处理,提出了一种改进的快速独立成分分析(FastICA)算法,采用基于加权相关系数进行白化处理以减少信息损失,并优化牛顿迭代法使其满足三阶收敛。对算法进行了细致描述,分析了算法的时间复杂度。实验结果表明,该方法可有效减少数据信息损失,具有迭代次数少、收敛速度快等优点,可有效提高入侵检测样本数据的优化效率。

关键词: 入侵检测, 快速独立成分分析, 数据优化, 牛顿迭代法

Abstract:

For the purpose of achieving the better data optimizat sing results in intrusion detection, an improved FastICA algorithm was proposed. The weighted correlation coefficient was adopted in the phase of albinism processing to reduce information loss, and the Newton's iterative method was improved for third-order convergence. The algorithm was introduced concretely, meanwhile the time complexity was analyzed in detail. The experiment shows that the method has the advantages of less times of iteration and fast speed of convergence, which can effectively decrease the losses of data and increase the efficiency of data optimization in in ion detection.

Key words: intrusion detection, fast independent component analysis, data optimization, Newton's iteration method

No Suggested Reading articles found!