Network intrusion detection in high dimensional space

Author

Marcek, Stanislav ; Drozda, Martin ; Juhas, Gabriel ; Lehocki, Fedor

Author_Institution

Fac. of Electr. Eng. & Inf. Technol., Slovak Univ. of Technol., Bratislava, Slovakia

fYear

2009

fDate

24-27 Nov. 2009

Firstpage

1

Lastpage

7

Abstract

We propose and evaluate an approach for network intrusion detection in high dimensional space. This approach is based on an approximate solution to the nearest neighbor problem. Our evaluation is based on the KDD´99 data set, a Yahoo Web spam data set and another set used in the NIPS´03 feature selection challenge. The approximate approach shows that good performance in terms of detection rate and false positives rate can be achieved.

Keywords

Internet; search engines; security of data; unsolicited e-mail; KDD´99 data set; NIPS´03 feature selection; Yahoo Web spam data set; high dimensional space; nearest neighbor problem; network intrusion detection; Computational modeling; Computer science; Computer simulation; Information systems; Information technology; Intrusion detection; Nearest neighbor searches; Space technology; Viruses (medical); Web and internet services;

fLanguage

English

Publisher

ieee

Conference_Titel

Applied Sciences in Biomedical and Communication Technologies, 2009. ISABEL 2009. 2nd International Symposium on

Conference_Location

Bratislava

Print_ISBN

978-1-4244-4640-7

Electronic_ISBN

978-1-4244-4641-4

Type

conf

DOI

10.1109/ISABEL.2009.5373652

Filename

5373652