Seminář: Anomaly detection

Nowadays, anomaly (outlier, rare event) detection is a very active research and application area. After a short introduction to anomaly detection and its succesful applications, I bring an overview of three settings – supervised, semi-supervised and unsupervised, and introduce basic methods for detecting anomalies. Then I will focus on current research directions, namely on evaluation of anomaly detection methods, on anomaly explanation, class-based outliers and outliers in text.