Algoritmo de células dendríticas através do modelo do perigo para um sistema de detecção de intrusos baseado em anomalias

Abstract

Intrusion detection systems are techniques to find if a network is having unauthorized access that can be an anomaly, caused by wrong software configuration or an inappropriate use or hacker attack. This systems are able to analyze network traffic, some activities occurring in a computer and to decide if the situations are simple events or some hacker attack. The reason to propose this final paper is due to the problems found about current classification algorithms. Intrusion detection systems have some problems like false positive, false negative, subversion errors and physical constraints on infrastructure. These problems are responsible to cause a large error of rate decision. This final paper propose a implementation of an algorithm supported by artificial intelligence techniques to decrease the error rate by intrusion detection algorithms.