Recommending in Social Tagging Systems based on Kernelized Multiway Analysis ⋆ Alexandros Nanopoulos and Artus Krohn-Grimberghe Institute of Computer Science, Information Systems and Machine Learning Lab University of Hildesheim, Germany {nanopoulos,artus}@ismll.de Summary. Along with the new opportunities introduced by Web 2.0 and collabora- tive tagging systems, several challenges have to be addressed too, notably the prob- lem of information overload. Recommender systems are among the most successful approaches for increasing the level of relevant content over the “noise”. Traditional recommender systems fail to address the requirements presented in collaborative tagging systems. This paper considers the problem of item recommendation in col- laborative tagging systems. It proposes to model data from collaborative tagging systems with 3-mode tensors, in order to capture the 3-way correlations between users, tags, and items. By applying multi-way analysis, latent semantic correlations are revealed, which help to improve the quality of recommendations. Nevertheless, high-order tensors tend be sparse, a fact that hinders the application of multi-way analysis. To address this problem, we propose the application of kernel-based meth- ods, which act as smoothing functions against sparsity. Experimental comparison, using data from a real collaborative tagging system (Bibsonomy), indicates the su- periority of the proposed method against the non kernel-based method and also against other baseline methods. Key words: Social tagging, Recommender systems, Multi-way analysis, Kernel functions. 1 Introduction Social tagging (a.k.a. collaborative tagging) is a process by which users assign labels in the form of keywords to a set of resources with a purpose to share, discover and recover them. Discovery enables users to find new content of their interest, that is shared by other users. Nowadays, collaborative tagging ⋆ The authors gratefully acknowledge the partial co-funding of their work through the European Commission FP7 project MyMedia (www.mymediaproject.org) under the grant agreement no. 215006. For your inquiries please contact info@mymediaproject.org.