pandas - How to efficiently do data augmentation in movieLen case for recommender systems in python?

Question

I have a dataset that is basically like the MovieLen dataset. It has 3 columns: userGUID, card, action. User GUID is userId, card is contentId, action is ratings of contents by users. It looks like as below:

I have 16536 users and 16734 cards And only little portion of them is rated (action). What I want to do is to is data augmentation where I get userGUID and card rows which are not rated at all.

So I would ultimately like to have each user's all card that is not rated (action-ed'). Each user will have augmented data excluding already rated ones. So data I expect would look like below:

This table is only for not rated augmented ones.

I wish someone can provide python code to do this.

Thank you in advance.

Answer 1

等待大神答复