A method inspired from differential coefficient for calculating danger signals in Artificial Immune System

Artificial Immune System is a computational system which is inspired from vertebrate immune system. For the characteristics of adaptability and diversity, Artificial Immune System is used to solve engineering problems, especially computer security problems. Danger Model is a new research focus of Artificial Immune System which is inspired from innate immune system. Previous researches have proved that Danger Model is effective in intrusion detection. The calculation of danger signals is the kernel of Danger Model, which is not effectively solved. In this paper we considered that danger comes from changes. Staring from monitoring the changes of computer systems, inspired from the principle of differential calculus, a calculating method based on changes is presented. Preliminary experiment has proved that such a danger signal calculating method based on changes can detect a certain kind of Trojan in Windows XP successfully, without any priori knowledge.