Precision on holdout data

Spot the Trend
1

Hello,

In the problem overview page we read:

However, be cautious of overfitting as the holdout data for this challenge will be more diverse than what is in the train and test set. One of the goals of this challenge is to create a generalizable model that can be deployed in many different data conditions.

Could you tell us a bit more about the holdout data? In which sense it is more diverse?

Thanks

2

Hi @konstantina.kontoudi

Great question, the holdout data is a larger dataset that has more variability in dynamics of the time series.

Onward Team