resample provides a set of tools for performing randomization-based inference in Python, primarily through the use of bootstrapping methods and Monte Carlo permutation tests. See the example notebook for a brief tutorial.
- Bootstrap samples (ordinary or balanced, both with optional stratification) of arrays with arbitrary dimension
- Parametric bootstrap samples (Gaussian, Poisson, gamma, etc.) of one-dimensional arrays
- Bootstrap confidence intervals (percentile, BCA and Studentized) for any well-defined parameter
- Randomization-based variants of traditional statistical tests (t-test, ANOVA F-test, K-S test, etc.)
- Tools for working with empirical distributions (empirical cumulative distribution and quantile functions, distance metrics for comparing distributions)
Installation requires numpy and scipy.
The latest release can be installed from PyPI:
pip install resample