Neural networks with categorical valued inputs and applications to intrusion detection

Author

Alexander Novokhodko

Keywords and Phrases

AdaBoost

Abstract

"The main contribution of this study is a neural network intrusion detector, built with a specific goal to minimize false alarms. The obtained performance, 0% false positive rate and very low 0.09% false negative rate, is superior to other published results. Additionally, it has been proven that Hamming distance can be successfully utilized with Gaussian kernels in probabilistic neural networks for categorical valued input data"--Introduction, page 1.

Advisor(s)

Wunsch, Donald C.

Committee Member(s)

Miller, Ann K.
Acar, Levent
Stanley, R. Joe
Dagli, Cihan H., 1949-

Department(s)

Electrical and Computer Engineering

Degree Name

Ph. D. in Computer Engineering

Comments

Partial support for this research from the National Science Foundation, from Sandia National Laboratories, and from the M. K. Finley Missouri endowment, is gratefully acknowledged.

Publisher

University of Missouri--Rolla

Publication Date

Fall 2004

Journal article titles appearing in thesis/dissertation

• Host based intrusion detection using Probabilistic Neural Networks and AdaBoost
• Probabilistic neural networks with Hamming distance kernels for categorical inputs

Pagination

ix, 55 pages

Note about bibliography

Includes bibliographical references.

Rights

© 2004 Alexander Novokhodko, All rights reserved.

Document Type

Dissertation - Citation

File Type

text

Language

English

Subject Headings

Computer securityNeural networks (Computer science)Radial basis functions

Thesis Number

T 8645

Print OCLC #

62253239

Recommended Citation

Novokhodko, Alexander, "Neural networks with categorical valued inputs and applications to intrusion detection" (2004). Doctoral Dissertations. 1594.
https://scholarsmine.mst.edu/doctoral_dissertations/1594