Bootstrapping (statistics)

Small sample sizes can mislead standard error estimates

Imagine you're tasting a new ice cream flavor for the first time and only have one scoop to judge its quality.

With just one scoop, you can't tell if it's good or bad; you need more scoops to get a better sense of the flavor's true quality. This is like needing more data to accurately estimate the standard error.

Example

You taste one scoop and say it's good, but you don't know if it would be good or bad if you had more scoops to taste.

Remember this

Just like you need multiple scoops to judge the ice cream, you need more data to accurately estimate the standard error.

Text adapted from Wikipedia, licensed under CC BY-SA 4.0.

Related concepts

Resampling (statistics)

Bootstrapping samples with replacement to estimate distributions

Global Forest Change dataset

Global Forest Change dataset covers 2000-2024

Boosting (machine learning)

Boosting reduces bias in ML models

the back-door criterion identifies: sufficient adjustment sets for causal estimation

Can we trust studies without random experiments?

to standardize: when you need zero mean and unit variance for gradient-based optimization

Why do we need to make data uniform before training a model?

Bias vs variance: high bias = underfitting, high variance = overfitting

Can a perfect fit to past data predict future events?

Swipe through 100 ML concepts daily

Open Pocket Polymath