macrOData : New Benchmarks of Thousands of Datasets for Tabular Outlier Detection
Xueying Ding, Simon Klüttermann, Haomin Wen, Yilong Chen, Leman Akoglu · 2026
Read macrOData : New Benchmarks of Thousands of Datasets for Tabular Outlier Detection by Xueying Ding, Simon Klüttermann, Haomin Wen, Yilong Chen, Leman Akoglu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-09329.txt.
Description
Outlier detection on tabular data underpins fraud checks, monitoring, and many other applied tasks, yet its leading benchmark covers only a few dozen datasets. Xueying Ding and four co-authors answer with MacrOData, a 2026 suite of nearly two and a half thousand curated datasets built for fair, statistically solid comparison of detection methods. Scale is the point: small suites cannot show how methods differ across varied data.
The suite has three parts. One gathers hundreds of datasets with real semantic anomalies. A second gathers hundreds more with real statistical outliers, and a third adds hundreds of synthetic datasets spanning varied data priors and outlier types. Standard train and test splits, public and private partitions with held-out labels, and semantic annotations come with the release. Trials across classical, deep, and foundation models yield reference scores and practical guidance, and the suite is open with a public leaderboard.
Catalog record
| CREATOR | Xueying Ding, Simon Klüttermann, Haomin Wen, Yilong Chen, Leman Akoglu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 102 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-09329