Back to Search Start Over

Uses and Misuses of Bayes' Rule and Bayesian Classifiers in Cybersecurity.

Authors :
Bard, Gregory V.
Source :
AIP Conference Proceedings; 2017, Vol. 1910 Issue 1, p1-8, 8p
Publication Year :
2017

Abstract

This paper will discuss the applications of Bayes' Rule and Bayesian Classifiers in Cybersecurity. While the most elementary form of Bayes' rule occurs in undergraduate coursework, there are more complicated forms as well. As an extended example, Bayesian spam filtering is explored, and is in many ways the most triumphant accomplishment of Bayesian reasoning in computer science, as nearly everyone with an email address has a spam folder. Bayesian Classifiers have also been responsible significant cybersecurity research results; yet, because they are not part of the standard curriculum, few in the mathematics or information-technology communities have seen the exact definitions, requirements, and proofs that comprise the subject. Moreover, numerous errors have been made by researchers (described in this paper), due to some mathematical misunderstandings dealing with conditional independence, or other badly chosen assumptions. Finally, to provide instructors and researchers with real-world examples, 25 published cybersecurity papers that use Bayesian reasoning are given, with 2-4 sentence summaries of the focus and contributions of each paper. [ABSTRACT FROM AUTHOR]

Details

Language :
English
ISSN :
0094243X
Volume :
1910
Issue :
1
Database :
Complementary Index
Journal :
AIP Conference Proceedings
Publication Type :
Conference
Accession number :
126716202
Full Text :
https://doi.org/10.1063/1.5013945