Bootstrapping a confidence interval in Python

Resampling gives an interval for any statistic, with no formula needed.

import numpy as np
from scipy import stats

rng = np.random.default_rng(5)
sample = rng.normal(95, 12, size=60)

result = stats.bootstrap((sample,), np.median, n_resamples=2000,
                         confidence_level=0.95, random_state=rng)
low, high = result.confidence_interval
print(round(np.median(sample), 2), round(low, 2), round(high, 2))
print(round(result.standard_error, 3))

perm = stats.permutation_test((sample, sample + 3),
                              lambda a, b: np.mean(a) - np.mean(b),
                              n_resamples=500, random_state=rng)
print(round(perm.pvalue, 4))

How it works

1. The data goes in as a one-element tuple of samples.
2. statistic is any function reducing a sample to a number.
3. confidence_level and n_resamples control the result.

Keywords and builtins used here

• as
• lambda
• print
• round

The run, in numbers

Lines

16

Characters to type

517

Tokens

157

Three-star pace

105 tpm

At the three-star pace of 105 tokens a minute, this run takes about 90 seconds.

Type this snippet

Step 2 of 2 in Correlation & resampling, step 9 of 23 in Scientific computing with SciPy.

← Previous Next →