benedekrozemberczki / shapley Goto Github PK

Hi Benedek,

Congrats and thanks for your great work for the community!

I compared the v1 and v2 versions of the paper on arxiv. I found that some content in sec.4.2 about how to measure ensemble heterogeneity using Shapley values was removed in the v2.

I wonder if I can use the Shapley values calculated by Troupe to measure the diversity or heterogeneity of an ensemble?
And is there any principle for measuring?

Best,

Mxk

Thank you for sharing your great work. I truly enjoyed reading it. However, I met an error when I tried the example. It seems to be fine for the MC example.

$ python multilinear_extension_example.py
RuntimeWarning: invalid value encountered in true_divide
self._Phi = self._Phi / np.sum(self._Phi, axis=1).reshape(-1, 1)
Traceback (most recent call last):
File "multilinear_extension_example.py", line 11, in
solver.solve_game(W, q)
File "/lib/python3.6/site-packages/shapley/solvers/multilinear_extension.py", line 34, in solve_game
self._run_sanity_check(W, self._Phi)
File "/lib/python3.6/site-packages/shapley/solution_concept.py", line 28, in _run_sanity_check
self._verify_distribution(Phi)
File "/lib/python3.6/site-packages/shapley/solution_concept.py", line 22, in _verify_distribution
assert np.sum(Phi) - Phi.shape[0] < 0.001
AssertionError

Hi, thank you for the nice repo.

I am a bit confused about the value in W. Could you use a classification example to give me some guidance?

Thank you for the help.