Hello, the text has been updated. Thank you for the good catch!

There are 4 200 customers, across five countries over a period of seven weeks. The training set is the first 7 weeks of data, and the last day and final week will be the test set, where the target is whether a client performed event 14 during that day at 6-hour intervals.

The evaluation metric for this challenge is F1.

I hope this clears up the confusion and good luck!

Thanks so much

i see, thanks for your reply.

ok, so our time-series prediction models will be evaluated on activities happened in only '1 DAY'

hmmm,

any reasons why you chose this evaluation strategy?

In my understanding, evaluation period should be more than at least 7 days...

by the way,

what does 'the last day and final week will be the test set' mean...?

are you going to add extra 1 week to current test file?

Hmmmmm, this changes the competition completely. Until before this explanation, we had to predict the transactions, now we have to predict if a given transaction took place.

@zindi so the submission file target column, is that now a 0 or 1 based on if this specific transaction took place?