What is the price of approximation? The saddlepoint approximation to a likelihood function
Godrick Oketch, Rachel M. Fewster, Jesse Goodman · 2025
Read What is the price of approximation? The saddlepoint approximation to a likelihood function by Godrick Oketch, Rachel M. Fewster, Jesse Goodman (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-2504-07324.txt.
Description
Godrick Oketch, Rachel Fewster, and Jesse Goodman study the cost of an approximation in statistics in two thousand twenty-five. The saddlepoint approximation can stand in for a likelihood when the true one is too hard to compute. But the estimate from the approximation will differ somewhat from the true maximum likelihood estimate.
Past work sized that gap by simulation, or by tackling the hard likelihood anyway. The authors instead give an explicit formula that can be computed directly. Readers can use it to judge whether the saddlepoint estimate is close enough. They test the formula in cases where the true likelihood is known. They also give asymptotic results that show how the gap behaves in a suitable limit.
Catalog record
| CREATOR | Godrick Oketch, Rachel M. Fewster, Jesse Goodman |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 142 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-07324