Various Aspects of Recommender Systems

Projects

The projects offered here are aligned with the lecture "Data Analysis and Query Languages" offered in the SS16, which covered some of the main RSs techniques.

This page will be continuously updated

- Cross-domain recommendations in RecSesame

Source: Recommender Systems Handbook. Francesco Ricci, Lior Rokach, Bracha Shapira.

Short description

Cross-domain recommender systems (CDRS) aim at generating recommendations that span across multiple domains by transferring knowledge from a source to a target domain, with the assumption that there are strong dependencies between items of those domains. For instance one can recommend movies to a user based on the information of the music she listens to. Traditional approaches show limitations when dealing with this problem.

The goal of this project is to integrate a recommender algorithm into RecSesame, a framework developed at out department, which generates cross-domain recommendations. The efficacy of the algorithm can be evaluated by the platform itself by means of an off-line evaluation.

Pre-requisites: knowledge of RDF, recommender systems. Programming in Java, Python; use of Maven.

Recommended number of participants: 2 per group

Status: assigned (2 groups of 2 students each)