What is the difference between recommender systems and prediction systems?

Question

data mining

I just can’t understand what is the difference between recommender systems and prediction?
Let's take the example of a store. If a person bought milk, then with a probability of 90% he will buy bread. This is a prediction. With it, you can line up products on the shelves so that a person buys related products. But after all, this is the task of recommender systems .. Who can explain?

Answer 1

Recommendations are based on collaborative filtering (usually, but not always) - it is simply a search in the history of completed actions by users similar to you.
Predictions, on the other hand, are based on process modeling (usually, but not always), and based on this model, a forecast is calculated (which may never have been seen in history).
Although, of course, no one forbids you to use predictions for recommendations and vice versa.