Visit our website: http://bit.ly/2GtXaiw
Bingjie Wang, Software Engineer, Oscar Health
Abstract
Recommending high quality and low cost providers to members is a critical part of Oscar’s mission to make healthcare easy and affordable. Our recommendations are powered by data from historical claims, member interactions, and direct scheduling, among other sources. This talk dives into how Oscar builds its rankings, starting from creating a custom objective function based on claims conversions (validated visits from our recommendations). This objective is then optimized with learn-to-rank techniques. Finally, analytical measurements, data pipelines, and feedback loops ensure that the recommendations further our mission of improving healthcare.
Add comment