Anomaly-detection methods are aimed at identifying observations that deviate manifestly from what is expected. Such methods are usually run on low-dimensional data, such as time series data. However, the increasing importance of high-dimensional payments and exposure data for financial oversight...