BDBComp
Parceria:
SBC
Detecção de Intrusos usando Conjunto de k-NN gerado por Subespaços Aleatórios

Márcia HenkeCelso CostaEulanda M. dos SantosEduardo Souto

Several studies have been proposed in the literature to deal with Internet anomaly detection by using machine learning techniques. Most of these works use individual classifiers such as k-NN (k-Nearest Neighbor), SVM (Support Vector Machines), Artificial Neural Networks, Decision Tree, Naive Bayes, k-means, among others. However, the literature has recently focused on applying classifier combination in order to increase detection rate. In this paper, a set of classifiers, more precisely, a set of k-NN generated throughRandom Subspaces Method is employed. Such an ensemble of classifiers method is compared to the hybrid technique TANN (Triangle Area based Nearest Neighbor), published recently in the literature. Results obtained using ensemble of k-NNs were superior to those obtained with TANN in terms of classification accuracy as well as false alarm reduction rate.

http://www.lbd.dcc.ufmg.br/colecoes/sbseg/2011/002.pdf

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato: bdbcomp@lbd.dcc.ufmg.br
     Mantida por:
LBD