Super Learner

Super Learner is a targeted machine learning algorithm developed by University of California Berkeley professor Mark van der Laan and Harvard University professor Sherri Rose. It allows for utilization of ensemble machine learning algorithms in the fitting of probability distributions on observational data towards a target parameter defining a scientific question of interest. The book Targeted Learning builds upon foundational work in the field of casual inference done by Judea Pearl.



Implementation in R

The SuperLearner prediction method is implemented in a CRAN package freely available for the open source R programming language.