TU Wien:Recommender Systems VU (Neidhardt)/Exam 2022-06-28

This is only a selection of questions, there is a large pool of possible questions.

• Recall the notion of F1-score of a recommender system. Given a precision of 0.27 and a recall of 0.46, what is the F1-score of the recommender?

• Recall the notion of recall of a recommender system. Let us assume that in a movie catalogue there are 16 movies that are relevant for user u. What is the recall of the top-10 recommendations if 2 are relevant for user u?

• Recall the notion of precision of a recommender system. Let us assume that in a movie catalogue there are 20 movies that are relevant for user u. What is the recall of the top-18 recommendations if 6 are relevant for user u?

• Recall the notion of Mean Absolute Error MAE. Compute MAE for a test set of the following predictions where the true rating is always 5 but the system predicts in 72 cases 5, in 26 cases 4, and in 11 cases 3.

• Calculate parts of item-item similarity (according to the formula from the lecture)

Materialien

Neues Material hinzufügen

B

• Bild2.png (details)
• Bild5.png (details)
• Bild6.png (details)