Discriminating Self from NonSelf with Finite Mixtures of Multivariate Bernoulli Distributions
Affinity functions are the core components in negative selection to discriminate self from nonself. It has been shown that affinity functions such as the rcontiguous distance and the Hamming distance are limited applicable for discrimination problems such as anomaly detection. We propose to model self as a discrete probability distribution specified by finite mixtures of multivariate Bernoulli distributions. As byproduct one also obtains information of nonself and hence is able to discriminate with probabilities self from nonself. We underpin our proposal with a comparative study between the two affinity functions and the probabilistic discrimination.
Discriminating Self from NonSelf with Finite Mixtures of Multivariate Bernoulli Distributions
Proceedings of the Genetic and Evolutionary Computation Conference (GECCO)
Authors:  Thomas Stibor 
Year/month:  2008/ 
Booktitle:  Proceedings of the Genetic and Evolutionary Computation Conference (GECCO) 
Publisher:  ACM 
Fulltext: 
Abstract 

Affinity functions are the core components in negative selection to discriminate self from nonself. It has been shown that affinity functions such as the rcontiguous distance and the Hamming distance are limited applicable for discrimination problems such as anomaly detection. We propose to model self as a discrete probability distribution specified by finite mixtures of multivariate Bernoulli distributions. As byproduct one also obtains information of nonself and hence is able to discriminate with probabilities self from nonself. We underpin our proposal with a comparative study between the two affinity functions and the probabilistic discrimination. 
Bibtex:
@inproceedings {author = { Thomas Stibor},
title = { Discriminating Self from NonSelf with Finite Mixtures of Multivariate Bernoulli Distributions },
year = { 2008 },
booktitle = { Proceedings of the Genetic and Evolutionary Computation Conference (GECCO) },
publisher = { ACM },
}