> Definition of COVID-19 positive vs negative


In the provided Jupyter notebook, there is a cell that explains:
"For this example cohort building notebook, we will label imaging studies as "COVID-positive" (or "COVID-negative", respectively) if a positive (or negative, respectively) COVID test result was obtained up to 14-days prior to the imaging study. So, we'll filter the DataFrame of studies for a `days_from_test_to_study` in the range of 0 to 14 days. The most recent test prior to the imaging study (within the time window) determines the imaging study COVID status label."

Is this the same definition of COVID-19 positive and COVID-19 negative that will be used for the competition? Or, is the definition being used for classification not public? I don't believe I've seen this explained elsewhere.

Thank you

Posted by: morgandaly @ Sept. 22, 2022, 8:55 p.m.

thank you for your question. Yes, the same COVID-positivity criterion was used to create the validation and test sets (from unpublished data).

Posted by: kdrukker @ Sept. 22, 2022, 8:58 p.m.

Great, thank you for the quick reply!

Posted by: morgandaly @ Sept. 22, 2022, 9 p.m.
Post in this thread