On the self representation based on the adaptive self radius

Self representation problem is very important for the negative selection algorithms. In this paper, two methods, named as K-NN-ASR and B-NN-ASR, are proposed to calculate the self radius of a self sample in the continuous space. As for the K-NN-ASR, the self radius of a self sample is calculated according to the distances from itself to the surrounding self samples. In the B-NN-ASR, both the position of the self sample and the distances to the surrounding self samples are considered. Experiments are done to test the performance of the proposed methods. Experimental results demonstrate that the proposed methods are competitive, and the B-NN-ASR is better.