An immune inspired model for obfuscated virus detection

Computer virus scanner is a vital approach to deal with computer virus. However, current static scanning techniques for virus detection have serious limitations. Motivated by a recent success in Computer Immune Theory and N-gram text classification method, an immune inspired obfuscated virus detection model is presented, which is referred as IOVDM. In IOVDM, N-gram analysis is applied to automatically generate gene lib from virus files, then generate immature cells from the gene lib, if the negative selection are succeed, they become mature cells. The mature cells will evolved into memory cells if they received co-stimulation. Finally, both memory and mature cells are used for classification. We use IOVDM for detection of unseen and obfuscated virus; we compared the detection ability of our model with three most prevalent anti-virus software. Favorable experimental results are obtained and presented.