search

clinicalml / cfrnet

Counterfactual Regression
MIT License
295 stars 82 forks source link

There are not 747 observations in ihdp_npci_1-1000.train.npz #7

Closed jameszhou-gl closed 3 years ago

jameszhou-gl commented 3 years ago

Hi, thanks very much for your code and paper. I find there are 672 observations(125 treatment, 547 control) in ihdp_npci_1-1000.train/test.npz. But the paper says 747 observations(139 treatment, 608 control).

jameszhou-gl commented 3 years ago

Got the problem. The original data is split to trian/test.