Re-ranking Recommendations Based on Predicted Short-term Interests – A Protocol and First Experiment

The recommendation of additional shopping items that are potentially interesting for the customer has become a stan-dard feature of modern online stores. In academia, research on recommender systems (RS) is mostly centered around ap-proaches that rely on explicit item ratings and long-term user profiles. In practical environments, however, such rating in-formation is often very sparse and for a large fraction of the users very little is known about their preferences. Fur-thermore, in particular when the shop offers products from a variety of categories, the decision of what should be recom-mended can strongly depend on the user’s current short-term interests and the navigational context. In this paper, we report the results of an initial experimental analysis evaluating the predictive accuracy of different con-textualized and non-contextualized recommendation strate-gies and discuss the question of appropriate experimental de-signs for such types of evaluations. To that purpose, we intro-duce a parameterizable protocol that supports session-specific accuracy measurements. Our analysis, which was based on log data obtained from a large online retailer for clothing and lifestyle products, shows that even a comparably simple con-textual post-processing approach based on product features can leverage short-term user interests to increase the accuracy of the recommendations