Bootstrapping in statistics
Bootstrapping is a method used to estimate variance, accuracy, and other metrics of sample estimates, such as the arithmetic mean. The simplest bootstrapping procedure consists of the following steps:
- Generate a large number of samples from the original data sample having the same size
N
. You can think of the original data as a jar containing numbers. We create the new samples byN
times randomly picking a number from the jar. Each time we return the number into the jar, so a number can occur multiple times in a generated sample. - With the new samples, we calculate the statistical estimate under investigation for each sample (for example, the arithmetic mean). This gives us a sample of possible values for the estimator. ...
Get NumPy : Beginner's Guide - Third Edition now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.