Evidence theory data fusion-based method for cyber-attack detection

Abstract

Detecting electronic crimes is a big challenge as they are tainted with a large number of imperfections such as imprecision and uncertainty. For this reason, we must choose a perfect tool to detect cybercrime taking place in an uncertain environment. In this paper, we are proposing a high-level data fusion approach based on evidence theory (Dempster-Shafer theory) which aims at improving the reliability of cybercrimes detection process using a more improved decision consisting in merging complementary decisions from two independent classifiers, namely Support Vector Machine (SVM) and Artificial Neural network (ANN) in an uncertain environment. In fact, the retained approach is characterized by its capability to overcome the uncertain data nature.