THE BOOTSTRAP

preview_player
Показать описание
The bootstrap is a computational technique to estimate the standard error of an estimator. We explain the procedure and analyze it for the sample mean. We then apply it to build confidence intervals for the correlation coefficient using a real-world dataset.

Photo by Ivan Bolshakov on Unsplash
Рекомендации по теме
Комментарии
Автор

In the case of a small sample size, can the batch size of bootstrap samples exceed the number of empirical samples? What is a good heuristic for determining batch size with respect to sample size in general?

StephenSpivack