Anomaly Detection-Based Negative Selection Algorithm

The real-valued negative selection algorithm and binary-valued negative selection algorithm are an anomaly detection approach, motivated by the negative selection immune system principle. In the training stage, solely positive (normal) samples are obtainable of some anomaly detection applications. Further, during the training phase traditional classification algorithms require samples for total classes. This chapter begins with a review of anomaly detection. Then, we present a description of negative selection algorithms.