a7b23 / h-divergence Goto Github PK

Hi, thanks for sharing the code. I was reading the code and it seems that the agriculture example you are using a H-Min divergence instead of the H-JS divergence mentioned in the paper for this particular example. Is the code different than the one used in the paper or was it a typo in the paper ? Thank you !

For sampling s1, s2 in the blobs experiments, the same seed is being used np.random.seed(seed=11 * k + 10 + n) in each trial. Ref.

Hey

I was wondering when you guys will upload the code for the H-Div sample quality evaluation given in the appendix of your paper.

Thanks

Hi,
Thank you for sharing the code.
I have a query regarding threshold computation, for e.g. in line 166 of utils.py.
Shouldn't S_jsv_vector[np.int(np.ceil(N_per * (1 - alpha)))] instead be S_jsv_vector[int(N_per * (1 - alpha))]?
Especially when N_per is small (say 3), ceil(N_per * (1 - alpha)) can be equal to N_per with the default alpha, causing indexing error.
Please let me know if I am missing something.

Regards,
Piyushi