jknife

Framework-agnostic Jackknife estimation utilities for statistical modeling

Overview

jknife provides a generic implementation of the jackknife resampling method for statistical computing. It is designed to be framework-agnostic, making it compatible with scikit-learn, statsmodels, and custom models.

Features

• Leave-one-out jackknife estimation with optional parallelization
• Support for scikit-learn estimators via adapter functions
• Intuitive API designed for both novice and expert users
• Minimal dependencies (only numpy and scipy for core functionality)

Installation

# Basic installation
pip install jknife

# With scikit-learn support
pip install jknife[sklearn]

# With parallel processing support
pip install jknife[parallel]

# With all optional dependencies
pip install jknife[all]

# For development
pip install jknife[dev]

Quick Example

import numpy as np
from sklearn.linear_model import LinearRegression
from sklearn.datasets import make_regression
from jknife.core import jackknife
from jknife.contrib.sklearn_adapters import sklearn_fit_fn, sklearn_coef_fn

# Generate some data
X, y = make_regression(n_samples=200, n_features=5, noise=10, random_state=42)

# Perform jackknife estimation
result = jackknife(
    X,
    y,
    fit_fn=sklearn_fit_fn(LinearRegression, fit_intercept=True),
    coef_fn=sklearn_coef_fn,
    n_jobs=-1,  # Use all available cores
)

# Print summary table
print(result.summary())

License

UNLICENSE