Font Size: 

In this talk, I will present Statistically Enhanced Learning (SEL), a general feature extraction approach to improve any type of learning technique, be it statistical or machine learning. Through the addition of highly informative covariates, which are obtained as statistical estimates rather than directly observed, SEL improves model learning for any type of data (tabular, computer vision, text). I will discuss the general idea, refer to existing feature extraction methods which actually can be shown to fall under the umbrella of SEL, and illustrate its performance on both simulated and real data. In particular, we will see how SEL allows improved predictions of handball tournaments and discuss how it can be used to derive a metric for teams' or players' strength.