Scribd
Upload
2 views

Bootstrapping Methods

Uploaded by

Priyanka Naik
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
2 views

Bootstrapping Methods

Uploaded by

Priyanka Naik
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 4

Bootstrapping Methods

Bootstrapping
• In statistics and machine learning, bootstrapping is a
resampling technique that involves repeatedly drawing
samples from our source data with replacement, often
to estimate a population parameter.
• By “with replacement”, we mean that the same data
point may be included in our resampled dataset
multiple times.
• Typically our source data is only a small sample of the
ground truth. Bootstrapping is loosely based on the law
of large numbers, which says that with enough data the
empirical distribution will be a good approximation of
the true distribution.
• Using bootstrapping, we can generate a distribution of
estimates, rather than a single point estimate. The
distribution gives us information about certainty, or the
lack of it.
Let’s use the same approach to calculate a confidence
interval when evaluating the accuracy of a model on a held-
out test set. Steps:
• Draw a sample of size N from the original dataset with
replacement. This is a bootstrap sample.
• Repeat step 1 S times, so that we have S bootstrap samples.
• Estimate our value on each of the bootstrap samples, so that
we have S estimates
• Use the distribution of estimates for inference (for example,
estimating the confidence intervals).

You might also like