MetaLearned Algorithm Selection for Recommender Systems
Dr. Jöran Beel, Trinity College Dublin
Event details  

What 

When 
Sep 04, 2019 from 15:00 to 16:00 
Where  SeminarRaum 2.025 
Add event to calendar 
vCal iCal 
Selecting and configuring the most effective algorithm for recommender systems is a nontrivial task as even experts struggle to predict, which algorithms may potentially perform well in a given system. We tackle the algorithm selection problem in recommender systems from a metalearning perspective, in which we learn and predict a) the overall best algorithm for a given problem space, and b) the best algorithm for each instance in the problem space.