thanks Debs.

I believe the months and years in the datasets are likely the total number of months or years that the users have been discussing or submitting @iq test

The test file is for validation, it's the same for everyone. The only difference it has from the training data is that the variables on the train set aids the decision of if a user is active or not, in my opinion. And yes you may get errors at confusion matrix level, there a way around it, but other than looking at accuracy, precision/AUC helps you answer the question.

Madam Deborah could you help me understand why we are given a zip file containing birthday user challenge when the aim of the competition is to predict who will be active the following month? Am a bit lost on that one.

Hello Isaac, I'm also in this challenge as you, but if you read through the description and datasets properly, I'm very sure you'll find an answer to your question.

Thanks Deby. I got it sorted out already but just wanted to be sure am dealing with the right dataset. If you don't mind we can make a team together, a winning team.

Which datasets have you examined and where exactly are you confused @astute?