A comparison of negative and positive selection algorithms in novel pattern detection

This paper describes a technique based on immunological principles for novel (anomalous) pattern detection. It is a probabilistic method that uses a negative selection scheme (complement pattern space) to detect any changes in the normal behavior of monitored data patterns. The technique is compared with a positive selection approach (implemented by an ART neural network), which uses the (self-) pattern space for anomaly detection. Some experimental results in both cases are reported.