05_bootstrap.rst

Bootstrap Confidence Intervals

Estimate parameter uncertainty.

Basic Usage

from distfit_pro.core.bootstrap import Bootstrap
import numpy as np
from distfit_pro import get_distribution

data = np.random.normal(10, 2, 1000)
dist = get_distribution('normal')
dist.fit(data)

# Parametric bootstrap
ci = Bootstrap.parametric(data, dist, n_bootstrap=1000)

for param, result in ci.items():
    print(result)

Non-Parametric Bootstrap

# Resample from data
ci = Bootstrap.nonparametric(data, dist, n_bootstrap=1000)

Interpretation

If 95% CI is [9.8, 10.2], we are 95% confident the true mean is in this range.