QAP examples

This directory contains public-safe examples for QAP models.

Current examples:

python examples/qap/qap_logit_toy.py
python examples/qap/qap_logit_backend_comparison.py

The toy examples use a synthetic directed binary network and two dyadic predictors:

• same_group
• distance

Backend options

QAPLogit supports two logistic-regression backends:

• backend="statsmodels": default backend with richer diagnostics;
• backend="sklearn": faster regularized backend for larger permutation runs.

The scikit-learn backend does not provide standard errors, z-values, or model-based backend p-values. QAP empirical p-values are still computed from the permutation distribution.

The output should be interpreted as a workflow demonstration, not as a benchmark or substantive social-science claim.