Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Min-max hyperellipsoidal clustering for anomaly detection in network security.

Suseela T Sarasamma1, Qiuming A Zhu

  • 1Northrop Grumman Mission Systems, Bellevue, NE 68023, USA. suseela_ts@yahoo.com

IEEE Transactions on Systems, Man, and Cybernetics. Part B, Cybernetics : a Publication of the IEEE Systems, Man, and Cybernetics Society
|August 15, 2006
PubMed
Summary

A new hyperellipsoidal clustering method enhances network intrusion detection. This technique improves anomaly detection accuracy, identifying specific threats missed by other systems.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Hierarchical Kohonenen net for anomaly detection in network security.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Societyยท2005
See all related articles

Area of Science:

  • Computer Science
  • Network Security
  • Machine Learning

Background:

  • Network security relies on effective intrusion detection systems (IDS) to identify malicious activities.
  • Existing anomaly detection methods struggle with identifying specific, novel attack patterns.
  • Accurate data modeling is crucial for robust intrusion detection.

Purpose of the Study:

  • To introduce a novel hyperellipsoidal clustering technique for network intrusion detection.
  • To enhance the capability of detecting individual anomaly types that are challenging for other schemes.
  • To identify optimal feature subsets for improved detection performance.

Main Methods:

  • Developed a hyperellipsoidal clustering technique generating clusters with high intracluster similarity and low intercluster similarity.

Related Experiment Videos

  • Utilized accretive processes to incrementally derive parameters for higher-order data models (multivariate Gaussian functions).
  • Implemented the technique within a feedforward neural network using Gaussian radial basis functions and evaluated using sample inclusiveness/exclusiveness criteria.

    • Main Results:

    • The hyperellipsoidal clustering technique demonstrated effectiveness in network intrusion detection.
    • Achieved above 95% detection rates with false-positive rates below 5% on tcptrace network-connection records.
    • Identified specific feature subsets that significantly improve detection accuracy.

    Conclusions:

    • The novel hyperellipsoidal clustering technique offers a significant advancement in intrusion detection systems.
    • This method excels at detecting individual anomaly types, improving overall network security.
    • The approach provides a robust framework for building more accurate and efficient anomaly detection models.