Looking for the dataset csuite_fork_collider_nonlin_gauss_latent_confounder

[fred887]

Hello,

I would like to reproduce the results of the D-DECI method (from v0.0.0 version) on the synthetic data in the corresponding paper.\ But I cannot find the csuite_fork_collider_nonlin_gauss_latent_confounder dataset.\ Neither here in https://github.com/microsoft/causica nor in https://github.com/microsoft/csuite.

Could you please indicate me where I can get this dataset or maybe provide me some source code to generate it by myself?

Thank you very much for your help,

Bests,

Frederic

[KaushikSathvara]

@fred887 Did you get any chance to find dataset? csuite_fork_collider_nonlin_gauss_latent_confounder

[fred887]

Unfortunately, no I could no get the dataset. I generated my own datasets using the source code in causica/large_synthetic/data_degeneration.py

[KaushikSathvara]

Yeah, I tried that as well. But seems it's throwing me some runtime error like below.

RuntimeError: mat1 and mat2 must have the same dtype

[LaurantChao]

Hi Frederic,

Thanks for reaching out.

I have now updated the csuite data to include csuite_fork_collider_nonlin_gauss_latent_confounder, which should be accessible (please refer to microsoft/csuite: CSuite: A Suite of Benchmark Datasets for Causality (github.com).

Best, Chao

[fred887]

Hi Chao, Thank you very much! Best, Frederic