Analysis Group Sponsors a Conference on Machine Learning Econometrics at the 36th Meeting of the Canadian Econometric Study Group

November 22, 2019

Analysis Group sponsored, participated in, and presented an award at the 36th Meeting of the Canadian Econometric Study Group, held at the Université du Québec à Montréal (UQAM). The conference, with the theme of “Machine Learning Econometrics,” was organized by the Departments of Economics at UQAM and Concordia University.

At the conference, Analysis Group consultants showcased a selection of the firm’s data science capabilities. Associate Iain Snoddy gave a presentation titled “Learning about Selection: An Improved Correction Procedure,” in which he discussed his work developing a methodology that uses machine learning (ML) tools to overcome selection bias in high-dimensional settings. Dr. Snoddy highlighted key differences between his method and the most widely used existing solutions, and noted a wide range of factors to which it can be applied to reduce the risk of selection bias, such as education, place of residence, workplace, occupation, and contract type.

The opening poster session included a presentation of “Estimating Average Treatment Effects with Propensity Scores Estimated With Four Machine Learning Procedures: Simulation Results in High Dimensional Settings With Time to Event Outcomes,” a paper by Principal Jimmy Royer, academic affiliate Professor Phil Merrigan of UQAM, and Analysis Group alumnus Kip Brown. Their paper evaluates four ML methods that have particular utility in the analysis of ever-growing sources of international big data related to health economics and outcomes research (HEOR).

During the conference, Analysis Group’s Data Science and Statistical Modeling practice group presented an award for the Best Student Poster on Machine Learning to Ph.D. candidate Harold D. Chiang from Vanderbilt University. His poster was titled “Many Average Partial Effects in L1-Regularized Binomial and Fractional Regressions: With an Application to Gendered Language on the Internet.”