Hi @trberg and L3C support,
As the data is time-sensitive, I would like to know how the final test set is split. Is it a random split which means the test set has similar time distribution to the training dataset, or the test dataset is with late time compared to the training dataset?
Best
Fanyou
Created by Fanyou Wu Fanyou_Wu Hi @Fanyou_Wu,
The training data will stay the same for the main quantitative evaluation. We may change it around though in additional experiments to see how your submission performs with varying inputs.
Thanks,
Tim Thanks @trberg,
So how about the training set? Is the training set be fixed, or will there be new training data? In the second scenario, we can still consider the spitting is random even though there are some new data.
Best
Fanyou Hi @Fanyou_Wu,
The current hold out test set was a random split from the training data. However, we are also including any long covid patients that accumulated over the course of this challenge. So to answer your question....both.
Thanks!
Tim
Drop files to upload
How the final test set are split? page is loading…