How do I correctly run the statistical comparison?

[cavalab]

How do I correctly run the statistical comparison?

[lacava]

Hi, can you be more specific? Are you not able to reproduce this notebook? Note it requires running the black-box and ground truth notebooks first to produce summary rankings for the stats comparisons. I described in issue #41 how we handle failed runs for the statistical comparisons. Note that for those comparisons there should be very few missing summary results, and only for AIFeynman. If you're getting a different result from the notebook, please provide details so I can assist.

[hengzhe-zhang]

Okay, thanks for your response. It's true that I should run the black box notebook in advance. I didn't realize that the notebook requires a file where it comes from the running result of the black box notebook. Thank you for your response. I have solved my problem.

[lacava]

I wanted to mention in commit 63637df we added collate scripts to post-processing, so the gap between the .json results and saved .feather results is eliminated.