1
$\begingroup$

I implemented a "normalized" confidence rule (see here: enter image description here) according to a course from coursera https://www.coursera.org/learn/recommender-systems) in pandas (see here: https://d2.maxfile.ro/knwouzvjhg.html) but I am unable to find any additional information on the formula that the instructor provided on the net. Can anyone help me out? Is this something similar to association rule mining?

How can I evaluate the performance of such a recommender?

$\endgroup$
0
$\begingroup$

Evaluation of recommender systems can be done by splitting the entire dataset into Training and Test datasets. Metrics like precision, recall, ROC can be used for binary ratings and metrics like RMSE, MAE can be used for absolute ratings.

These error metrics can be analyzed across split time dimensions, different user clusters to note if there are any seasonal pattern or built in clusters within the data

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.