Semiparametric KSD test: unifying score and distance-based approaches for goodness-of-fit testing
Zhihan Huang, Ziang NiuDepartment of Statistics and Data ScienceUniversity of Pennsylvania · 2025
Read Semiparametric KSD test: unifying score and distance-based approaches for goodness-of-fit testing by Zhihan Huang, Ziang NiuDepartment of Statistics and Data ScienceUniversity of Pennsylvania (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-2512-20007.txt.
Description
Zhihan Huang and Ziang Niu make goodness-of-fit testing cheap enough to use on awkward models. Many statistical models have normalizing constants nobody can compute, so checking whether one actually fits the data is hard: sampling from the fitted model is expensive, and bootstrap checks demand repeated refitting. Their semiparametric test builds on the kernelized Stein discrepancy, which can judge fit using only the model's score function.
They establish the statistic's consistency and its asymptotic null distribution under general nuisance estimation, then propose an influence-adjusted wild bootstrap that produces a valid test without refitting or fresh sampling. The bootstrap's consistency and local power are proven, and simulations from normality testing to intractable likelihoods show competitive power at a fraction of rivals' cost. A lung-cancer protein signaling network provides the applied illustration.
Catalog record
| CREATOR | Zhihan Huang, Ziang NiuDepartment of Statistics and Data ScienceUniversity of Pennsylvania |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 189 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2512-20007