skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Statistical foundations of audit trail analysis for the detection of computer misuse

Journal Article · · IEEE (Institute of Electrical and Electronics Engineers) Transactions on Software Engineering; (United States)
DOI:https://doi.org/10.1109/32.241771· OSTI ID:5636136
 [1];  [2]
  1. Univ. of New Mexico, Albuquerque, NM (United States). Computer Science Dept.
  2. Oak Ridge National Lab., Oak Ridge, TN (United States) Univ. of Tennessee, Knoxville, TN (United States). Computer Science Dept.

The authors model computer transactions as generated by two stationary stochastic processes, the legitimate (normal) process N and the misuse process M. They define misuse (anomaly) detection to be the identification of transactions most likely to have been generated by M. They formally demonstrate that the accuracy of misuse detectors is bounded by a function of the difference of the densities of the processes N and M over the space of transactions. In practice, detection accuracy can be far below this bound, and generally improves with increasing sample size of historical (training) data. Careful selection of transaction attributes also can improve detection accuracy; they suggest several criteria for attribute selection, including adequate sampling rate and separation between models. They demonstrate that exactly optimizing even the simplest of these criteria is NP-hard, thus motivating a heuristic approach. They further differentiate between modeling (density estimation) and nonmodeling approaches. They introduce a frequentist method as a special case of the former, and Wisdom and Sense, developed at Los Alamos National Laboratory, as a special case of the latter. For nonmodeling approaches such as Wisdom and Sense that generate statistical rules, they show that the rules must be maximally specific to ensure consistency with Bayesian analysis. Finally, they provide suggestions for testing detection systems and present limited test results using Wisdom and Sense and the frequentist approach.

OSTI ID:
5636136
Journal Information:
IEEE (Institute of Electrical and Electronics Engineers) Transactions on Software Engineering; (United States), Vol. 19:9; ISSN 0098-5589
Country of Publication:
United States
Language:
English