KGMS10b

Wael Khreich, Eric Granger, Ali Miri, Robert Sabourin. Boolean Combination of Classifiers in the ROC Space. In 20th International Conference on Pattern Recognition, Pages 4299-4303, Los Alamitos, CA, USA, 2010.

Abstract

Using Boolean AND and OR functions to combine the responses of multiple one- or two-class classifiers in the ROC space may significantly improve performance of a detec- tion system over a single best classifier. However, techniques found in literature assume that the classifiers are conditionally- independent, and that their ROC curves are convex. These assumptions are not valid in most real-world applications, where classifiers are designed using limited and imbalanced training data. A new Iterative Boolean Combination (IBC) technique applies all Boolean functions to combine the ROC curves produced by multiple classifiers without prior assump- tions, and its time complexity is linear according to the number of classifiers. The results of computer simulations conducted on synthetic and real-world host-based intrusion detection data indicate that combining the responses from multiple HMMs with IBC can achieve a significantly higher level of performance than with the AND and OR combinations, especially when training data is limited and imbalanced

Contact

W. Khreich
E. Granger
Ali Miri
R. Sabourin

BibTex Reference

@InProceedings{KGMS10b,
   Author = {Khreich, Wael and Granger, Eric and Miri, Ali and Sabourin, Robert},
   Title = {Boolean Combination of Classifiers in the ROC Space},
   Journal = {Pattern Recognition, International Conference on},
   BookTitle = {20th International Conference on Pattern Recognition},
   Pages = {4299--4303},
   Publisher = {IEEE Computer Society},
   Address = {Los Alamitos, CA, USA},
   Year = {2010}
}

EndNote Reference [help]

Get EndNote Reference (.ref)