Neuro-Immune and Self-Organizing Map Approaches to Anomaly Detection: A Comparison

The purpose of this work is to investigate a hybrid approach (neuro-immune technique) for anomaly detection on time series data. In many anomaly detection applications, only positive (normal) samples are available for training purpose. However, conventional classification algorithms need both positive and negative samples. The proposed approach uses normal samples to generate abnormal samples that are subsequently used as training data for a neural network. The approach is compared against an anomaly detection technique that uses self-organizing maps to cluster the normal data sets (samples).1