Differences between our training dataset and your (Challenge Team's) training dataset

[Elaine Li]

Dear Challenge Team,

How will the organizers handle the differences between our training dataset choice and the training dataset used in the evaluation by the organizers?

For example, If we choose not to use some CODE-15% data in my training dataset, should we inform you in the README?

Will you run both the uploaded model and train our model using your training dataset? If so, which score will be used for the competition?

Thanks,

Best regards,

Elaine

[PhysioNet Challenge]

Dear Elaine,

Each of the records identifies its source, so your training code can skip some of the data sources or some of the records in a source. You can mention your training process in your README, and you should describe your training process in your paper, but the important part is implementing your training process in your training code.

We will train your model using your training code on the training set, and we will evaluate the resulting model on the validation set.

Best,
Matt
(On behalf of the Challenge team.)

Please post questions and comments in the forum. However, if your question reveals information about your entry, then please email info at physionetchallenge.org. We may post parts of our reply publicly if we feel that all Challengers should benefit from it. We will not answer emails about the Challenge to any other address. This email is maintained by a group. Please do not email us individually.