I just wonder that how I should define the examples if I’m trying to build collaborative filtering model where targets are clicks instead of ratings? For example in e-commerce I want to recommend products that people might like based on the products they have watched. If I just target to be 1 everytime user clicks meaning that I add user id and item id to database with 1 is it going to work? Should I also add cases where user didn’t pressed the item? In my use case it’s really hard to determine when user didn’t pressed the item but still understood what it was and that’s why I’m interested of knowing that will the model work if there is only examples of user item pairs where target value is 1?