Jerome H. Friedman
Stanford University
|
Primary Section: 32, Applied Mathematical Sciences Membership Type:
Member
(elected 2010)
|
Research Interests
As a statistician my research has predominantly centered on the statistical machine learning problem. The goal is to use data derived from some system under study to try to learn or understand its nature, and/or to predict aspects of its behavior in the future. My focus has been on the development of general purpose learning machines, implemented in software, that can be used in a wide variety of settings with different types of data. Primarily based on the data itself, with minimal user input or guidance, these programs attempt to automatically extract relevant patterns in the data characterizing intrinsic properties of the system that can then be used to make reliable inferences and predictions concerning its properties. Over the years I have developed or been involved in developing a variety of machine learning techniques that have been successfully applied to data from a wide variety of scientific disciplines, as well as medical, commercial, and industrial applications. These include classification and regression trees (CART), multivariate adaptive regression splines (MARS), multiple additive regression trees (MART) and rule based predictive ensembles (RuleFit).
