An Immune Concentration Based Virus Detection Approach Using Particle Swarm Optimization

This paper proposes an immune concentration based virus detection approach which utilizes a two-element concentration vector to construct the feature. In this approach, ‘self’ and ‘nonself’ concentrations are extracted through ‘self’ and ‘nonself’ detector libraries, respectively, to form a vector with two elements of concentrations for characterizing the program efficiently and fast. Several classifiers including k-nearest neighbor (KNN), RBF neural network and support vector machine (SVM) with this vector as input are then employed to classify the programs. The selection of detector library determinant and parameters associated with a certain classifier is here considered as an optimization problem aiming at maximizing the accuracy of classification. A clonal particle swarm optimization (CPSO) algorithm is used for this purpose. Experimental results demonstrate that the proposed approach not only has a very much fast speed but also gives around 98% of accuracy under optimum conditions.