Craig Scott
Senior Product Manager
Elsevier
Martin Rajman
Head of the SCI-IC-MR Research Group
École Polytechnique Fédérale de Lausanne
This briefing will address the École Polytechnique Fédérale de Lausanne (EPFL)-Elsevier partnership that has delivered a new recommendation engine for the ScienceDirect.com website. The new recommendation engine makes use of collaborative filtering techniques, exploiting available usage data consisting of approximately five billion article download events. The resource was developed using a minimum viable product (MVP) methodology, taking a performance metric-driven approach with multiple iterations measured through A/B testing; it leveraged high performance computing cluster (HPCC) distributed computing in a private cloud, and ultimately delivered an impressive increase in usage of the feature. The talk will also address the planned next phase of the EPFL-Elsevier partnership, which aims to create more targeted recommendations, customized for either individual researchers or researchers belonging to a specific segment, e.g. a given subject field or career stage.