Hi,
I’m trying to complete this competition, in this moment I tested my model with training data separated in 2 parts. However, I’d like to test with test labels, in order to know the performance with validation data.
Could anyone help me with this issue?
Best regards,
Santiago Arboleda
Hi, im testing this example with the official datasets downloaded here:
https://www.drivendata.org/competitions/66/flu-shot-learning/data/
Hi @sarboledaq. The test labels are withheld because predictions on the test set are used for leaderboard scoring. You can get a score on your test set predictions by submitting your predictions to the competition: https://www.drivendata.org/competitions/66/flu-shot-learning/submissions/
Splitting the training data (like what you’re already doing) is the intended way for you to do your own development validation.