fix(stats): keep regression inference alive when squared deviations underflow

This commit is contained in:
2026-09-27 19:26:04 +02:00
parent 275de79123
commit 32e2c1efab
3 changed files with 296 additions and 6 deletions
+4
View File
@@ -33,6 +33,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
p-values and `NoncentralTCDF` at extreme t answer the tail the
format still holds instead of a silent zero or an error naming a
NaN.
- `LinearRegression` and `WeightedLinearRegression` keep their
inference alive when the squared deviations underflow to zero: the
standard errors, t and F statistics read factored sums of squares
instead of reporting infinity beside zero evidence.
## [1.0.0] - 2026-09-03