Actuarial Science and Financial Mathematics seminar: Federico Echenique

Joint seminar with Department of Economics

Federico Echenique
University of California-Berkeley

Room: M3 3127

Learning preferences

I will describe how to formulate the problem of recovering economic preferences and utility functions as a binary classification problem, which allows us to use standard ideas from machine learning in the recovery of these objects from choice data. The methods are broadly applicable in environments with noisy binary choice data. The focus of the talk will be on recent work leveraging choice data augmented with response time to recover preferences using recent developments in machine learning.