esda.fdr¶

esda.fdr(pvalues, alpha=0.05)[source]¶

Calculate the p-value cut-off to control for the false discovery rate (FDR) for multiple testing.

If by controlling for FDR, all of n null hypotheses are rejected, the conservative Bonferroni bound (alpha/n) is returned instead.

Parameters:¶

pvalues : array¶

(n, ), p values for n multiple tests.

alpha : float, optional¶

Significance level. Default is 0.05.

Returns:¶

Adjusted criterion for rejecting the null hypothesis. If by controlling for FDR, all of n null hypotheses are rejected, the conservative Bonferroni bound (alpha/n) is returned.

Return type:¶

float

Notes

For technical details see [Benjamini and Yekutieli, 2001] and [de Castro and Singer, 2006].

Examples

>>> import libpysal
>>> import numpy as np
>>> np.random.seed(10)
>>> w = libpysal.io.open(libpysal.examples.get_path("stl.gal")).read()
>>> f = libpysal.io.open(libpysal.examples.get_path("stl_hom.txt"))
>>> y = np.array(f.by_col['HR8893'])
>>> from esda import Moran_Local
>>> from esda import fdr
>>> lm = Moran_Local(
...     y,
...     w,
...     transformation="r",
...     permutations=999,
...     seed=12345,
...     alternative='two-sided',
... )
>>> fdr(lm.p_sim, 0.1)
0.001282051282051282

Return the conservative Bonferroni bound

>>> fdr(lm.p_sim, 0.05)
0.000641025641025641