Inspect values of derived (deterministic, un-transformed) variables at point?

junpenglao · May 25, 2019, 5:21am

There is this helper function you might find helpful:

pymc-devs/pymc3/blob/master/pymc3/util.py#L129-L144


def update_start_vals(a, b, model):
    """Update a with b, without overwriting existing keys. Values specified for
    transformed variables on the original scale are also transformed and inserted.
    """
    if model is not None:
        for free_RV in model.free_RVs:
            tname = free_RV.name
            for name in a:
                if is_transformed_name(tname) and get_untransformed_name(tname) == name:
                    transform_func = [
                        d.transformation for d in model.deterministics if d.name == name]
                    if transform_func:
                        b[tname] = transform_func[0].forward_val(
                            a[name], point=b)


    a.update({k: v for k, v in b.items() if k not in a})

I think if you do update_start_vals(testpoint, testpoint, model) it should output a dict with parameters that in the domain that the logp actually takes.

Topic		Replies	Views
Possible bug with backwards-transforming a BoundedContinuous variable v5	5	561	July 27, 2022
A general method to reversing transformations in custom step? Questions	0	454	May 24, 2021
Get the distribution of a Determinitic variable Questions	3	512	October 21, 2019
Get nan or inf from model.logp (model.test point) is an attestation of incorrectly configured model? Questions	22	8333	June 24, 2017
Cross-checking logp with scipy Questions	1	406	February 14, 2020

Inspect values of derived (deterministic, un-transformed) variables at point?

Related topics