hybridnonlinur

Nonlinear Hybrid Unit Root Tests — A comprehensive Python library for advanced econometric unit root testing.

Overview

hybridnonlinur implements 13 nonlinear unit root tests from the frontier of time series econometrics. Each test is faithfully translated from published papers and verified GAUSS source code.

Implemented Tests

Installation

pip install hybridnonlinur

Or from source:

git clone https://github.com/merwanroudane/hybridnonlinur.git
cd hybridnonlinur
pip install -e .

Quick Start

import numpy as np
import hybridnonlinur as hnl

# Load your time series
y = np.cumsum(np.random.randn(200))

# Run individual tests
result = hnl.fourier_kss(y, model=1)
print(result.summary())

# Run ALL 13 tests at once
results = hnl.run_all(y, series_name="My Series", model=1)

# Quantile unit root test
qr_result = hnl.qr_kss(y, model=1, taus=np.arange(0.1, 1.0, 0.1))
print(qr_result.summary())

# Bootstrap critical values
boot_result = hnl.fourier_kss_bootstrap(y, model=1, nboot=1000)
print(boot_result.summary())

Visualization

# Publication-quality comparison table
fig = hnl.plot_comparison_table(list(results.values()))
fig.savefig('comparison.png', dpi=300)

# Quantile profile plot
fig = hnl.plot_quantile_results(qr_result)
fig.savefig('quantile_profile.png', dpi=300)

# Series with Fourier fit
fig = hnl.plot_series_with_fourier(y, model=1, k=1)
fig.savefig('fourier_fit.png', dpi=300)

Model Specifications

• model=0: No deterministic components (QR-KSS only)
• model=1: Intercept only
• model=2: Intercept + trend

Test Categories

ESTAR Tests (t-statistic, reject if stat < CV)

KSS, Fourier-KSS, Fourier-ADF, KSS-FF, FWKSS

Kruse Tests (tau-statistic, reject if stat > CV)

Fourier-Kruse, FFKRUSE

AESTAR/Sollis Tests (F-statistic, reject if stat > CV)

Fourier-Sollis, FFSOLLIS, Fourier-Sollis (Zero-Mean)

Quantile Tests (t-statistic per quantile)

QR-KSS, QR-Fourier-KSS

Dependencies

• NumPy, SciPy, Matplotlib, tabulate

Author

Dr. Merwan Roudane
Email: merwanroudane920@gmail.com
GitHub: https://github.com/merwanroudane/hybridnonlinur

License

MIT License — see LICENSE for details.

Citation

If you use this library in your research, please cite:

@software{roudane2026hybridnonlinur,
  author = {Roudane, Merwan},
  title = {hybridnonlinur: Nonlinear Hybrid Unit Root Tests in Python},
  year = {2026},
  url = {https://github.com/merwanroudane/hybridnonlinur}
}