eigenstrapping.stats.gpd_inference
- eigenstrapping.stats.gpd_inference(perms, stat=0.15, per=0.1)[source]
Report the tail-estimated p-value for faster inference using a Generalized Pareto distribution.
- Parameters:
perms (np.ndarray of shape (N,)) – Statistical values with target brain map (e.g., Pearson correlation)
stat (float, optional) – Empirical stat (e.g. Pearson correlation) of original brain map and target brain map. Default is 0.15.
per (float, optional) – Fraction of highest correlation values to keep (0 < per <= 1). Default is 0.10.
- Returns:
Estimated p-value
- Return type: