News Articles Recommendation Algorithms

Abstract

In this work, we evaluate state-of-the-art recommender systems on the challenging task of recommending news articles. We develop novel recommendation and ranking approaches. More- over, we explore how the usage of the newest Natural Language Processing (NLP) techniques can affect the recommendation performance of the algorithms. On data sets of the Globo News Portal and Neue Züricher Zeitung we evaluate our methods against common baselines. Thereby, we measure a variety of metrics of the recommender system literature and include them in our analysis.