Hello,
Can anybody mention what the format of the predictions.csv file should be?
I don't understand it from the github file.
Thanks,
Ankita
Created by Ankita Roy ARTD @fmarotta ,
Great question! For the Final Round, you will only be able to submit one model to the queue. Hello @vchung ! Does this mean we can only make one submission, or are we allowed to submit multiple models, review their performance, and then select a final model for the final submission? @ARTD ,
Yes! You are more than welcome to continue working on your model as needed. However, please note that for the final round, you are only allowed to make one **total** submission (invalid submissions will not count towards this quota). Thanks! Also after the leaderboard submission end date can you still
upload the model?
Ankita @ARTD ,
`/output/predictions.csv` should be a 2-column CSV file with the following columns:
**Column Name** | **Column Type** | **Accepted Values**
--|--|--
`epr_number` | `str` | Sample/Participant IDs must be unique and match with those in the input files; there must be one prediction per sample ID or participant ID for PEGS participants.
`disease_probability` | `float` | All probabilities must be a number _between_ 0 (indicating no likelihood of the disease) and 1 (indicating 100% likelihood of having the disease); null/NaN values are not accepted
The expected output is also described in the [Submission Tutorial here](https://www.synapse.org/Synapse:syn52817032/wiki/627650).
Hope this helps!