|Table of Contents|

New SVM Weighted Feature Classification Method in Network Intrusion Detection

《南京理工大学学报》(自然科学版)[ISSN:1005-9830/CN:32-1397/N]

Issue:
2008年02期
Page:
231-236
Research Field:
Publishing date:

Info

Title:
New SVM Weighted Feature Classification Method in Network Intrusion Detection
Author(s):
ZHANG Kun12CAO Hong-xin2ZHONG Yi2LIU Feng-yu2
1.State Key Laboratory for Novel Software,Nanjing University,Nanjing 210093,China;2.School of Computer Science and Technology,NUST,Nanjing 210094,China
Keywords:
network intrusion detection support vector machine weighted feature classification
PACS:
TP393.08
DOI:
-
Abstract:
In view of the discovery that the different network data features have different influence on classification results in SVM-based network intrusion detection,a new SVM weighted feature classification method is brought forward in order to get optimal classification plane.The value of the data features that have more influence on classification is exponentially transformed by weighting.The falsely classified samples near classification plane are effectively corrected.The experimental results demonstrate that the proposed method can increase the number of samples that is classified rightly,and improve the accuracy of network intrusion detection.

References:

[1] 边肇祺, 张学工. 模式识别(第2版) [M ] . 北京: 清华大学出版社, 2000.
[2] Bonifac io J, C am sian A. Neu ra l netwo rks applied in intrusion detection system [ A ]. Proceedings o f The 1998 IEEE International Jo int Conference on Neu ra l Netwo rks [ C ]. P iscataw ay: The 1998 IEEE Internationa l Jo int Conference on Neural Netwo rks, 1998. 205- 210.
[3] Balajinath B, Raghavan S V. Intrusion detection through learning behav iormode l [ J]. ComputerCommun ications, 2001, 24( 12): 1 202- 1 212.
[4] Fo rrest S, Perre lason A S, A llen L, et a.l Sel-f nonself d iscrim ination in a compu ter [ A]. Proceedings of the 1994 IEEE Sym pos ium on Research in Security and Privacy [ C ] . Oakland: IEEE Computer Soc ie ty Press, 1994. 202- 212.
[5] 张琨, 徐永红, 王珩, 等. 基于免疫学的入侵检测 系统模型[ J]. 南京理工大学学报, 2002, 26( 4): 337- 340.
[6] 周莲英, 刘凤玉. 基于/ 群集智能0的入侵检测系 统研究[ J]. 南京理工大学学报, 2006, 30( 5): 587 - 591.
[7] 李千目, 戚湧, 张宏, 等. IIDS的行为特征提取方 法研究[ J]. 南京理工大学学报, 2004, 28( 2): 140 - 144.
[8] 张琨, 许满武, 刘凤玉, 等. 基于支持向量机的异 常入侵检测系统[ J]. 计算机工程, 2004, 30( 18): 43- 45.
[9] 李辉, 管晓宏, 昝鑫, 等. 基于支持向量机的网络 入侵检测[ J]. 计算机研究与发展, 2003, 40( 6): 799- 807.
[10] 饶鲜, 董春曦, 杨绍全. 基于支持向量机的入侵检 测系统[ J]. 软件学报, 2003, 14( 4): 798- 803.
[11] 陈光英, 张千里, 李星. 基于SVM 分类机的入侵检 测系统[ J]. 通信学报, 2002, 23( 5): 51- 56.
[12] 李昆仑, 黄厚宽, 田盛丰, 等. 模糊多类支持向量 机及其在入侵检测中的应用[ J]. 计算机学报, 2005, 28( 2): 274- 280.
[13] 李昆仑, 赵俊忠, 黄厚宽, 等. 基于SVM 技术的入 侵检测[ J]. 信息与控制, 2003, 32( 6) : 495- 498.
[14] 凌永发, 解季萍. 支持向量机在入侵检测系统中的 应用[ J]. 电力自动化设备, 2005, 25( 8): 59- 62.
[15] Info rma tion and Com puter Science, Un iversity o f C ali?? forn ia. KDD Cup 1999 Data [ EB /OL] . http: / /kdd. ics. uc .i edu /databases/kddcup99 /kddcup99. htm .l
[16] Lincoln Laboratory, M assachusetts Institute o f techno lo?? gy. DARPA Intrusion Detection Eva luation [ EB /OL]. http: / /www. l.l m it. edu / IST / ideva l/ index. htm .l
[17] 曹宏鑫. 基于SVM 的网络入侵检测研究[ D]. 南 京: 南京理工大学计算机科学与技术学院, 2004

Memo

Memo:
-
Last Update: 2008-04-30