Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Create Ordered Multinomial distribution #4773

Merged
merged 20 commits into from
Jul 5, 2021
Merged

Conversation

AlexAndorra
Copy link
Contributor

@AlexAndorra AlexAndorra commented Jun 15, 2021

This PR enables the ordered logistic constraint on multinomial observed data (i.e aggregated by trial).
Currently, OrderedLogistic only accepts data in a disaggregated format (i.e Categorical observed data).

I also added the option of having the inferred probabilities in the trace (not only the cutpoints, which are hard to interpret anyway). As I'm guessing that's mostly what people care about when using this likelihood, the option is True by default (it's only useful to disable it when memory usage is an issue).

This probably needs a few tests (inspired by those of OrderedLogistic), but here is already a proof of concept:

true_cum_p = np.array([0.1, 0.15, 0.25, 0.50, 0.65, 0.90, 1.0])
true_p = np.hstack([true_cum_p[0], true_cum_p[1:] - true_cum_p[:-1]])
fake_elections = np.random.multinomial(n=1_000, pvals=true_p, size=60)

with pm.Model() as m:
    cutpoints = pm.Normal(
        "cutpoints",
        mu=np.arange(6) - 2.5,
        sigma=1.5,
        initval=np.arange(6) - 2.5,
        transform=pm.distributions.transforms.ordered,
    )

    pm.OrderedMultinomial(
        "results",
        eta=0.0,
        cutpoints=cutpoints,
        n=fake_elections.sum(1),
        observed=fake_elections,
    )

    trace = pm.sample()

Which does recover the true probabilities:
complete_p

  • No breaking changes
  • New tests
  • Run linting / style checks
  • Mention in RELEASE-NOTES.md

pymc3/distributions/multivariate.py Outdated Show resolved Hide resolved
pymc3/distributions/multivariate.py Outdated Show resolved Hide resolved
pymc3/distributions/multivariate.py Show resolved Hide resolved
@ricardoV94
Copy link
Member

I also added the option of having the inferred probabilities in the trace (not only the cutpoints, which are hard to interpret anyway).

This would also be nice for the other Ordered distributions no?

@AlexAndorra
Copy link
Contributor Author

AlexAndorra commented Jun 16, 2021

This would also be nice for the other Ordered distributions no?

Just added that and addressed your other comments 🥳
Only the name issue and new tests remaining -- will handle this later

@ricardoV94
Copy link
Member

ricardoV94 commented Jun 16, 2021

Maybe it's easier to keep the two things separate. Have a helper class that creates the raw Ordered distribution and then, optionally, creates the deterministic p. You can retrieve p from the inputs that go into the variable returned from the raw Ordered distribution. Something along the lines of:

class _OrderedMultinomial(Multinomial):
  # standard implementation without the helper Deterministic

class OrderedMultinomial:
  def __new__(cls, name, *args, save_p=True, **kwargs):
    out_rv = _OrderedMultinomial(name, *args, **kwargs)
    if save_p:
      pm.Deterministic(f'{name}_p', out_rv.owner.inputs[4])  # I think this is the p vector in the multinomial
    return out_rv

  @classmethod
  def dist(cls, *args, **kwargs):
    return _OrderedMultinomial.dist(*args, **kwargs)

No need to even check if it's inside the model context, that's done by the default _OrderedMultinomial.

@ricardoV94
Copy link
Member

ricardoV94 commented Jun 16, 2021

Ignore the previous idea. I think you can overwrite __new__ to save name in self.name that's probably simpler. No, doesn't work either. The previous comment goes back to being my best idea :p

@AlexAndorra
Copy link
Contributor Author

Agreed, I was thinking the same -- that's actually how I did it for LKJCholeskyCov back in the days

@AlexAndorra
Copy link
Contributor Author

out_rv.owner.inputs[4]) does seem to work @ricardoV94 ! Small question though: how did you know that the p vector was in the 4th position? 😅

@ricardoV94
Copy link
Member

out_rv.owner.inputs[4]) does seem to work @ricardoV94 ! Small question though: how did you know that the p vector was in the 4th position? sweat_smile

The RVs inputs are always of the form (RandomStateSharedVariable, size, dtype, param1, param2 ..., paramN)

import aesara.tensor as at
x = at.random.multinomial(5, np.ones(3)/3, size=2)
print(x.owner.inputs)
[RandomStateSharedVariable(<RandomState(MT19937) at 0x7F125671EC40>),
 TensorConstant{(1,) of 2},
 TensorConstant{4},
 TensorConstant{5},
 TensorConstant{(3,) of 0...3333333333}]

@AlexAndorra
Copy link
Contributor Author

The RVs inputs are always of the form (RandomStateSharedVariable, size, dtype, param1, param2 ..., paramN)

Ooooh, ok, thanks! What does a dtype of TensorConstant{4} mean though? 🤔

@ricardoV94
Copy link
Member

ricardoV94 commented Jun 22, 2021

The RVs inputs are always of the form (RandomStateSharedVariable, size, dtype, param1, param2 ..., paramN)

Ooooh, ok, thanks! What does a dtype of TensorConstant{4} mean though? thinking

It's a numerical code for different dtypes ("float32", "float64", "int32", and so on). I don't know where they are defined though.

4 is probably for "int64" since that's the default dtype of discrete RVs

@AlexAndorra
Copy link
Contributor Author

Ah ok, I wouldn't have guessed that 😅
So the only thing left for this PR is adding a few tests. I think I need to write the logpdf for that but don't know how it works yet for this distribution

@ricardoV94
Copy link
Member

So the only thing left for this PR is adding a few tests. I think I need to write the logpdf for that but don't know how it works yet for this distribution

The logp shouldn't have to be tested as it's just the multinomial. The tests should just make sure the parameters that go into it are the correct ones. Have a look at the check_pymc_params_match_rv_op tests in test_random.py (guide), I think it does just what is needed.

Plus some tests for the helper Deterministic

@fonnesbeck fonnesbeck modified the milestones: vNext (4.0.0), v4.0.1 Jul 2, 2021
ricardoV94
ricardoV94 previously approved these changes Jul 4, 2021
Copy link
Member

@ricardoV94 ricardoV94 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM!

@ricardoV94 ricardoV94 dismissed their stale review July 4, 2021 06:06

Some conflicts in the files it seems. Other than that it looks good.

@codecov
Copy link

codecov bot commented Jul 4, 2021

Codecov Report

Merging #4773 (64b1ac8) into main (125256f) will increase coverage by 0.07%.
The diff coverage is 98.11%.

Impacted file tree graph

@@            Coverage Diff             @@
##             main    #4773      +/-   ##
==========================================
+ Coverage   71.97%   72.05%   +0.07%     
==========================================
  Files          85       85              
  Lines       13839    13872      +33     
==========================================
+ Hits         9961     9995      +34     
+ Misses       3878     3877       -1     
Impacted Files Coverage Δ
pymc3/distributions/__init__.py 100.00% <ø> (ø)
pymc3/distributions/multivariate.py 58.31% <95.45%> (+1.43%) ⬆️
pymc3/distributions/discrete.py 99.00% <100.00%> (+0.05%) ⬆️
pymc3/step_methods/hmc/nuts.py 96.87% <0.00%> (-0.63%) ⬇️

@ricardoV94
Copy link
Member

ricardoV94 commented Jul 4, 2021

The failing tests in distributions_random.py are due to the test looking for the rv_op which doesn't exist in the wrapper class. Probably just need to overwrite the relevant part in the inherited test subclasses.

@AlexAndorra
Copy link
Contributor Author

I just added back the rv_op in the wrapper class. Tests pass locally. Is that a good solution?

@ricardoV94
Copy link
Member

Is that a good solution?

Actually I think testing the wrapped distribution might the best. Just test the _Ordered* classes instead.

This shouldn't require any changes to the wrapper classes

@AlexAndorra
Copy link
Contributor Author

All tests are green 🥳
Just added a mention in RELEASE-NOTES.md

Copy link
Member

@ricardoV94 ricardoV94 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good! Left a minor suggestion that can be ignored.

pymc3/distributions/multivariate.py Outdated Show resolved Hide resolved
@ricardoV94
Copy link
Member

ricardoV94 commented Jul 5, 2021

The failing test comes from #4771. It can be ignored.

Finally, do we need to add an entry in the docs for the new distribution or is it automatic? I added it

@AlexAndorra
Copy link
Contributor Author

Merging 🍾 Thanks for your help and reviews @ricardoV94 !

@AlexAndorra AlexAndorra merged commit e5e83d0 into main Jul 5, 2021
@AlexAndorra AlexAndorra deleted the add-ordered-multinomial branch July 5, 2021 10:49
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants