Transfer learning for content-based recommender systems using tree matching

Abstract

In this paper we present a new approach to content-based transfer learning for solving the data sparsity problem in cases when the users' preferences in the target domain are either scarce or unavailable, but the necessary information for the preferences exists in another domain. Training a system to use such information across domains is shown to produce better performance. Specifically, we represent users' behavior patterns based on topological graph structures. Each behavior pattern represents the behavior of a set of users, when the users' behavior is defined as the items they rated and the items' rating values. In the next step, a correlation is found between behavior patterns in the source domain and target domain. This mapping is considered a bridge between the two. Based on the correlation and content-attributes of the items, a machine learning model is trained to predict users' ratings in the target domain. When our approach is compared to the popularity approach and KNN-cross-domain on a real world dataset, the results show that our approach outperforms both methods on an average of 83%. © 2013 IFIP International Federation for Information Processing.