around 500 user_id's data is not present in any csv files except users.csv.
Due to this no test data and features can be presented to predict this users's engagement.
True. I saw around 600 which is almost half. That realy affects the performance of the algos. Looks like it was intentiona, but why???
@Sorous @JuliusFx
You can view this challenge more of as a forecasting challenge. So the ~500 users in the user file are the users you need to "forecast" for and submit. You will find those ~500 users in the sample submission file.
True. I saw around 600 which is almost half. That realy affects the performance of the algos. Looks like it was intentiona, but why???
@Sorous @JuliusFx
You can view this challenge more of as a forecasting challenge. So the ~500 users in the user file are the users you need to "forecast" for and submit. You will find those ~500 users in the sample submission file.