As part of the Center for Social & Behavioral Science Methods Series, CSBS partners with the National Center for Supercomputing Applications (NCSA) to host the next session of the machine learning workshop led by Jeff Levy, Assistant Instructional Professor at the University of Chicago Harris School of Public Policy.
Building on the foundations introduced in the previous workshop, this session will focus on tree-based methods for classification and regression (e.g., decision trees, bagging, random forests, boosting). Participants will review the usage of tree-based methods in recent literature and apply these techniques in a hands-on exercise.
Recommended software includes Python and R, both freely available through Illinois Computes Research Notebooks.
Watch Session 1
In this session, Jeff Levy introduces tree-based machine learning methods for classification and regression. The presentation covers decision tree fundamentals, including recursive binary splitting and pruning, and explores techniques for improving model performance such as bagging, random forests, and boosting. The session highlights how these approaches can capture nonlinear relationships and complex interactions common in social and behavioral science data.
Watch Session 2
This session reviews recent applications of tree-based methods in social and behavioral science research and guides participants through implementing these models in R. Jeff Levy walks through example code and demonstrates how decision trees and related techniques can be applied to analyze complex datasets.
Watch Session 3
In this session, the Illinois Computes team introduces the no-cost computing resources available to researchers across the University of Illinois. They highlight access to advanced technologies, expert staff support, and tools designed for users at all levels of computing experience. The session also guides participants through the hands-on workshop component using Illinois Computes Research Notebooks, including environments such as Jupyter Notebooks and RStudio.
Instructor
Jeff Levy is an Assistant Instructional Professor at the University of Chicago Harris School of Public Policy. His background is in applied policy research, working at the intersection of social science and data science. His policy career started out at the Brookings Institution, followed by the data science and research programming team at the Urban Institute, where he worked on a wide range of projects on topics including state economies, non-profits, incarceration and the parole system. At Harris, Levy teaches both microeconomics, and the core sequence on programming and data skills. His own research has focused on economic uncertainty, particularly around the Great Recession.
He holds a BA in economics and political science from Michigan State University, and an MA and PhD in economics from American University in Washington, DC.
Contact Olivia Olvera (OliviaO@illinois.edu) with questions.