6

The Wikipedia article about Jacknife estimation of the bias and variance of an estimator $\theta$ includes the following formulas:

Variance of $\theta$:

$ \operatorname {Var}(\theta )=\sigma ^{2}={\frac {n-1}{n}}\sum _{{i=1}}^{n}({\bar {\theta }}_{i}-{\bar {\theta }}_{{\mathrm {Jack}}})^{2}$

where ${\bar {\theta }}_{{Jack}}={\frac {1}{n}}\sum _{{i=1}}^{n}({\bar {\theta }}_{i})$ is the jacknife estimator.

Bias-correction of $\theta$:

$ {\bar {\theta }}_{{\mathrm {BiasCorrected}}}=N{\bar {\theta }}-(N-1){\bar {\theta }}_{{Jack}} $

My question is: What are the corresponding formulas for bootstrap? Are they different?

Amelio Vazquez-Reina
  • 17,546
  • 26
  • 74
  • 110
  • 2
    Shao, Jun. “Bootstrap Variance and Bias Estimation in Linear Models.” The Canadian Journal of Statistics / La Revue Canadienne De Statistique, vol. 16, no. 4, 1988, pp. 371–382. JSTOR, JSTOR, www.jstor.org/stable/3314934. – Lawi Oscar Jul 03 '18 at 08:18

1 Answers1

2

I think the formulas for Bootstrap are as follows, although I can't seem to find a proper reference:

Let $h = \frac{1}{K} \sum_{k=1}^{K}{\hat{y}_k}$ be the mean predicted outcome from the $K$ bootstrap resamples.

Then:

  • Bias is $y - h$
  • Variance is $\frac{1}{K-1}\sum_{k=1}^{K} (y_k - h)^2$
oDDsKooL
  • 1,202
  • 2
  • 14
  • 32
  • Your bias is different from https://stats.stackexchange.com/questions/129478/when-is-the-bootstrap-estimate-of-bias-valid – user3226167 Apr 16 '19 at 09:52