Unsupervised Anomaly Detection Benchmark
Markus Goldstein · Harvard Dataverse · 2015
These datasets can be used for benchmarking unsupervised anomaly detection algorithms (for example "Local Outlier Factor" LOF). The datasets have been obtained from multiple sources and are mainly based on datasets originally used for supervised machine learning. By publishing these modifications, a comparison of different algorithms is now possible for unsupervised anomaly detection.