Evaluation of Adaptive Systems

Explore the intricacies of evaluating adaptive systems in this comprehensive tutorial from the UMAP'20 conference. Delve into group-based personalization models using topical user profiles and learn how they compare to individual user profile-based ranking methods. Discover techniques for resolving data sparsity issues, improving ranking results, and addressing the cold start problem for users without history. Examine the advantages of group profiles in terms of privacy and security. Gain insights into the effectiveness of short-term, long-term, and session-based user profiles, with a focus on improving Mean Reciprocal Rank and Normalized Discounted Cumulative Gain metrics. Led by experts Stephan Weibelzahl, Alexandros Paramythis, and Judith Masthoff, this 2-hour 18-minute session provides valuable knowledge for researchers and practitioners in user modeling, adaptation, and personalization.