MARS: Magnitude-Aware Rank Statistics Journal: Pattern Recognition Letters
Muhammad Rajabinasab, Afsaneh M. Nejad, Arthur Zimek · 2026
Read MARS: Magnitude-Aware Rank Statistics Journal: Pattern Recognition Letters by Muhammad Rajabinasab, Afsaneh M. Nejad, Arthur Zimek (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-2605-23563.txt.
Description
Rajabinasab, Nejad, and Zimek propose MARS, magnitude-aware rank statistics for comparing machine learning models. Standard difference diagrams rank models discretely, discarding how large gaps are, a flaw they call magnitude-blindness. MARS weights discrete ranks by a relative margin that scales with the distance between best and worst performers, with a dynamic projection for boundary cases. After a difference value is computed, MARS gives a truer picture of model differences and deeper insight into how methods truly perform across wide experiments.
Catalog record
| CREATOR | Muhammad Rajabinasab, Afsaneh M. Nejad, Arthur Zimek |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 30758 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.23563