pymc.model.transform.conditioning.change_value_transforms#

pymc.model.transform.conditioning.change_value_transforms(model, vars_to_transforms)[source]#

Change the value variables transforms in the model.

Parameters:

modelModel
vars_to_transformsDict: Dictionary that maps RVs to new transforms to be applied to the respective value variables

Returns:

new_modelModel: Model with the updated transformed value variables

Examples

Extract untransformed space Hessian after finding transformed space MAP

import pymc as pm
from pymc.distributions.transforms import logodds
from pymc.model.transform.conditioning import change_value_transforms

with pm.Model() as base_m:
    p = pm.Uniform("p", 0, 1, default_transform=None)
    w = pm.Binomial("w", n=9, p=p, observed=6)

with change_value_transforms(base_m, {"p": logodds}) as transformed_p:
    mean_q = pm.find_MAP()

with change_value_transforms(transformed_p, {"p": None}) as untransformed_p:
    new_p = untransformed_p["p"]
    std_q = ((1 / pm.find_hessian(mean_q, vars=[new_p])) ** 0.5)[0]

print(f"  Mean, Standard deviation\\np {mean_q['p']:.2}, {std_q[0]:.2}")
#   Mean, Standard deviation
# p 0.67, 0.16