Last active
March 19, 2024 17:23
-
-
Save JohannesBuchner/2027d0f313521387c2cded2424cdcfeb to your computer and use it in GitHub Desktop.
Highest Density Interval from posterior samples
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
from getdist.mcsamples import MCSamples | |
import getdist.chains | |
def highest_density_interval_from_samples(xsamples, xlo=None, xhi=None, probability_level=0.68): | |
""" | |
Compute the highest density interval (HDI) from posterior samples. | |
Parameters | |
---------- | |
xsamples : array_like | |
The posterior samples from which to compute the HDI. | |
xlo : float or None, optional | |
Lower boundary limiting the space. Default is None. | |
xhi : float or None, optional | |
Upper boundary limiting the space. Default is None. | |
probability_level : float, optional | |
The desired probability level for the HDI. Default is 0.68. | |
Returns | |
------- | |
tuple | |
A tuple containing the maximum a posteriori (MAP) estimate and the lower and upper | |
bounds of the HDI. | |
Notes | |
----- | |
The function starts at the highest density point and accumulates neighboring points | |
until the specified probability level is reached. If `xlo` or `xhi` is provided, | |
the HDI is constrained within these bounds. | |
Examples | |
-------- | |
>>> xsamples = np.random.normal(loc=0, scale=1, size=100000) | |
>>> hdi = highest_density_interval_from_samples(xsamples) | |
>>> print('x = %.1f + %.2f - %.2f' % hdi) | |
x = 0.0 + 1.02 - 0.96 | |
""" | |
getdist.chains.print_load_details = False | |
samples = MCSamples(samples=xsamples, names=['x'], | |
settings = {'mult_bias_correction_order':1}, | |
ranges={'x':[xlo,xhi]}) | |
density_bounded = samples.get1DDensityGridData('x') | |
x = density_bounded.x | |
y = density_bounded.P | |
# Sort the y values in descending order | |
sorted_indices = np.argsort(y)[::-1] | |
sorted_y = y[sorted_indices] / np.sum(y) | |
sorted_x = x[sorted_indices] | |
# Initialize variables | |
total_probability = sorted_y[0] | |
map = sorted_x[0] | |
x_lo = map | |
x_hi = map | |
for i in range(1, len(sorted_y)): | |
#print(total_probability, sorted_x[i]) | |
# Add the current probability to the total | |
total_probability += sorted_y[i] | |
x_lo = min(x_lo, sorted_x[i]) | |
x_hi = max(x_hi, sorted_x[i]) | |
# Check if the total probability exceeds or equals the desired level | |
if total_probability >= probability_level: | |
break | |
return map, map - x_lo, x_hi - map | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment