The Bootstrap is Inconsistent with Probability Theory
This paper proves that for no prior probability distribution does the bootstrap (BS) distribution equal the predictive distribution, for all Bernoulli trails of some fixed size. It then proves that for no prior will the BS give the same first two moments as the predictive distribution for all size trails. It ends with an investigation of whether the BS can get the variance correct.
Year of publication: |
1995-10
|
---|---|
Authors: | Wolpert, David H. |
Institutions: | Santa Fe Institute |
Saved in:
Saved in favorites
Similar items by person
-
Determining Whether Two Data Sets are From the Same Distribution
Wolpert, David H., (1995)
-
Wolpert, David H., (1995)
-
An Efficient Method to Estimate Bagging's Generalization Error
Wolpert, David H., (1996)
- More ...