Kernel Two-Sample Testing via Directional Components Analysis
Rui Cui, Yuhao Li, Xiaojun Song · 2025
Read Kernel Two-Sample Testing via Directional Components Analysis by Rui Cui, Yuhao Li, Xiaojun Song (2025) 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-2508-08564.txt.
Description
Rui Cui, Yuhao Li, and Xiaojun Song sharpen a test for whether two data sets come from the same source in a 2025 paper. The standard test sums gaps across all directions of a rich space, but the faint tail directions add noise and drain power. The new test keeps only the strong, well-read head directions.
That trim gives the test more punch, most of all with high bulk or lopsided groups. A fast fake-data scheme sets the pass mark, far quicker than shuffle-based ways. Trials on made and real data hold the false-alarm rate in line while beating the standard test's power.
Catalog record
| CREATOR | Rui Cui, Yuhao Li, Xiaojun Song |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 95 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-08564