Is it possible to have a better explanation of the expected labels?
For example, currently the data has months for creation date:
[ 1, 12, 2, 11, 3, 5, 4]
For each member with a creation date month equal to 5 are we meant to predict their activity chance in month 6? Why are there two months in the sample Submission?
I think people would compete more in this competition if working out what to predict wasn't so cryptic...
An easy way to help would be to simply provide a list of the IDs expected in the test set.
The target of this competition is ambigous.
Please can someone from Zindi clarify?
For example, is the current leaderboard only for people who joined in month 5 and predicting whether they are active in month 6?