Anomica: Fast Support Vector Based Novelty Detection

In this paper we propose ν-Anomica, a novel anomaly detection technique that can be trained on huge data sets with much reduced running time compared to the benchmark one-class...