---
id: hyppo
version: "0.5.2"
license: MIT License
license_treatment: permissive
maintenance: active
---

# hyppo — A comprehensive independence testing package

License: permissive · Maintenance: active · Downloads: 155.6K/mo

## What it is and what it does

hyppo is a Python library for multivariate hypothesis testing, designed to handle high-dimensional and nonlinear data where traditional univariate tests fall short. It implements a range of independence tests and other statistical procedures, including some novel methods not available in other packages. The library targets researchers and practitioners who need to test relationships between multiple variables simultaneously, filling a gap where R libraries exist but lack consistent interfaces or Python availability.

The package depends on core scientific Python tools (numpy, scipy, scikit-learn, statsmodels, numba, autograd, pandas, patsy, future) and is maintained actively. It runs on Python 3.8 through 3.14 and is in Alpha status, meaning the API may still evolve but the core functionality is usable for research and analysis work.

Use it for:

- Test independence between variables in high-dimensional datasets across multiple fields.
- Validate nonlinear relationships in multivariate data where standard correlation is insufficient.
- Compare statistical test results using multiple hypothesis testing methods on the same data.
- Conduct exploratory data analysis when you need to detect associations in complex, multi-variable systems.
- Implement novel independence tests documented in recent statistical literature not yet in mainstream packages.

## Worth the install?

AI-flagged interpretation of the facts above — verify before relying.

hyppo provides multivariate hypothesis testing for high-dimensional and nonlinear data, implementing both classical and novel statistical tests in Python.

Yes, if you need multivariate hypothesis testing for high-dimensional or nonlinear data. The package is actively maintained, has low install friction, carries a permissive MIT License, and fills a real gap in Python's statistical testing landscape. Alpha status indicates the API may evolve, so verify that the specific tests you need are documented and validated for your use case.

## Install

pip install hyppo
uv add hyppo
poetry add hyppo

## Installing hyppo

Before you install: Low friction installation with a pure-Python wheel. Active maintenance as of 2026-08-09 with recent commits. Depends on a substantial stack (numpy, scipy, scikit-learn, statsmodels, numba, autograd, pandas, patsy, future) but all are standard scientific Python packages.

License in practice: MIT License permits commercial and private use with minimal restrictions—you may use, modify, and distribute hyppo freely provided you include the license notice.

Quickstart:

pip install hyppo

from hyppo.independence import Hsic
import numpy as np

X = np.random.randn(10, 5)
Y = np.random.randn(10, 3)
stat, pval = Hsic().test(X, Y)

Verify before relying:

- Whether the package's novel tests are documented with references or citations to support their statistical validity.
- Performance characteristics and scalability limits for very high-dimensional datasets.
- Specific use-case suitability and validation status for genomics, neuroimaging, or other domains.

## Package facts

- License: MIT License (permissive)
- Python support: unspecified
- Install friction: low
- Maintenance: active
- Downloads: 155.6K/month (top 15,000 on PyPI)
- Known vulnerabilities: none known

## Tags

multivariate hypothesis testing, high dimensional statistical tests, independence testing python, nonlinear hypothesis testing, statistical significance testing, multivariate data analysis, hypothesis testing library, statistical-testing, multivariate-analysis, hypothesis-testing

[View on SkillFed](https://skillfed.io/packages/hyppo) · [View on PyPI](https://pypi.org/project/hyppo/)