stanford.seal64.gif (1,768 bytes)

Broad Area Colloquium for Artificial Intelligence,
Geometry, Graphics, Robotics and Vision

Predictive Data Mining with Multiple Additive Regression Trees

Jerome H. Friedman
Stanford University

Monday, April 15th, 2002, 4:45PM
Gates B01


Predicting future outcomes based on past observational data is a common application in data mining. The primary goal is usually predictive accuracy, with secondary goals being speed, ease of use, and interpretability of the resulting predictive model. New automated procedures for predictive data mining, based on "boosting" (CART) decision trees, are described. The goal is a class of fast "off-the-shelf" learning procedures that are competitive in accuracy with more customized approaches, while being fairly automatic to use (little tuning), and highly robust especially when applied to less than clean data. Tools are presented for interpreting and visualizing these multiple additive regression tree (MART) models.

About the Speaker

Jerome H. Friedman is a professor of statistics at Stanford University, and leader of the Computation Research Group at the Stanford Linear Accelerator Center. He recieved his bachelor and Ph.D degrees from the University of California at Berkeley, both in physics. His main research interests are in machine learning and data mining.

Back to the Colloquium Page