# How to Evaluate the Quality of Unsupervised Anomaly Detection Algorithms?

arXiv: https://arxiv.org/abs/1607.01152 · HTML: https://mlsanity.com/paper/1607.01152/

## Key facts

- Cited by 24 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 5 Jul 2016.

## Built on by (most cited first, 24 of 24)

- [Unsupervised Model Selection for Time-series Anomaly Detection](https://mlsanity.com/paper/2210.01078/) (2022)
- [Hyperparameter Sensitivity in Deep Outlier Detection: Analysis and a Scalable Hyper-Ensemble Solution](https://mlsanity.com/paper/2206.07647/) (2022)
- [Toward Unsupervised Outlier Model Selection](https://mlsanity.com/paper/2211.01834/) (2022)
- [Unsupervised Anomaly Detection with Rejection](https://mlsanity.com/paper/2305.13189/) (2023)
- [A Large-scale Study on Unsupervised Outlier Model Selection: Do Internal Strategies Suffice?](https://mlsanity.com/paper/2104.01422/) (2021)
- [Graph Anomaly Detection with Unsupervised GNNs](https://mlsanity.com/paper/2210.09535/) (2022)
- [Fast Unsupervised Deep Outlier Model Selection with Hypernetworks](https://mlsanity.com/paper/2307.10529/) (2023)
- [Meta-Learning for Unsupervised Outlier Detection with Optimal Transport](https://mlsanity.com/paper/2211.00372/) (2022)
- [Hyperparameter Optimization for Unsupervised Outlier Detection](https://mlsanity.com/paper/2208.11727/) (2022)
- [Leveraging the Christoffel Function for Outlier Detection in Data Streams](https://mlsanity.com/paper/2508.16617/) (2025)
- [Tax Evasion Risk Management Using a Hybrid Unsupervised Outlier Detection Method](https://mlsanity.com/paper/2103.01033/) (2021)
- [Is AUC the best measure for practical comparison of anomaly detectors?](https://mlsanity.com/paper/2305.04754/) (2023)
- [Self-Supervision for Tackling Unsupervised Anomaly Detection: Pitfalls and Opportunities](https://mlsanity.com/paper/2308.14380/) (2023)
- [Benchmarking Anomaly Detection Algorithms: Deep Learning and Beyond](https://mlsanity.com/paper/2402.07281/) (2024)
- [Towards Unsupervised Validation of Anomaly-Detection Models](https://mlsanity.com/paper/2410.14579/) (2024)
- [Unleashing the Potential of Unsupervised Deep Outlier Detection through Automated Training Stopping](https://mlsanity.com/paper/2305.16777/) (2023)
- [Deep Clustering Evaluation: How to Validate Internal Clustering Validation Measures](https://mlsanity.com/paper/2403.14830/) (2024)
- [Towards Automated Self-Supervised Learning for Truly Unsupervised Graph Anomaly Detection](https://mlsanity.com/paper/2501.14694/) (2025)
- [Automated Machine Learning for Unsupervised Tabular Tasks](https://mlsanity.com/paper/2510.07569/) (2025)
- [Analyzing Shapley Additive Explanations to Understand Anomaly Detection Algorithm Behaviors and Their Complementarity](https://mlsanity.com/paper/2602.00208/) (2026)
- [Automatic Unsupervised Ensemble Outlier Model Selection--Extended Version](https://mlsanity.com/paper/2605.16567/) (2026)
- [Self-Adaptive Anomaly Detection with Reinforcement Learning and Human Feedback in Connected Vehicles](https://mlsanity.com/paper/2607.08373/) (2026)
- [OneShot: Label-Efficient Detector Weighting for Time Series Anomaly Detection](https://mlsanity.com/paper/2610.16219/) (2026)
- [Retrieval Over Training: Similarity search-based Model Selection for Time Series Anomaly Detection](https://mlsanity.com/paper/2610.20243/) (2026)

## Abstract (arXiv)

When sufficient labeled data are available, classical criteria based on Receiver Operating Characteristic (ROC) or Precision-Recall (PR) curves can be used to compare the performance of un-supervised anomaly detection algorithms. However , in many situations, few or no data are labeled. This calls for alternative criteria one can compute on non-labeled data. In this paper, two criteria that do not require labels are empirically shown to discriminate accurately (w.r.t. ROC or PR based criteria) between algorithms. These criteria are based on existing Excess-Mass (EM) and Mass-Volume (MV) curves, which generally cannot be well estimated in large dimension. A methodology based on feature sub-sampling and aggregating is also described and tested, extending the use of these criteria to high-dimensional datasets and solving major drawbacks inherent to standard EM and MV curves.