Recommendations for Everyday Health

Students - Qiuya Xu, Xiaobo Shi, Ji Wu, Margaret Hall

Advisor - David Ruppert

Spring 2014

This past year, our MPS group partnered with Everyday Health, a leader in digital health and wellness.  Everyday Health reaches over 43 million users every month with data-driven, highly personalized content to help people live healthier lives.  We had the privilege of working with Everyday Health to provide a system that offered personalized article recommendations for anonymous users.

 Our team utilized several well-known predictive methods to approach the problem, such as decision trees and multinomial logistic regressions.  In addition, we worked with other recommender systems specialists to apply collaborative filtering methods.  We completed the project in SAS and R, utilizing our recently acquired skills from fall semester.  The training data from the final iteration of our project yielded an 80% accuracy rate.

 Our team was awarded one of two Best MPS Project awards for our varied statistical approaches and excellence in project execution.