Tag Archives: Ana

Research

SCOR Project Newsletter #2

1 Comment

The second newsletter, related to the SCOR research project is now available. It is a brief summary of the second six months block, from April till the end of September (the first one is available here).

As explained, over the past six months, we have had several interns, including Noé Bosc-Haddad, Florent Crouzet, Julien Siharath, Ana María Patrón Piñerez and Cassandra Mussard, the visit of Laurence Barry and Fei Huang, Philipp Ratz and Samuel Stocksieker defended their PhD, François Hu finished his postdoctoral fellowship, while Marouane Il Idrissi and Arsene Zotsa just arrived, Agathe Fernandes Machado and Olivier Côté (co-supervised with Ewen Gallic and Marie-Pier Côté) finished their PhD courses and are now 100% on their research… We wrote papers, gave talks… Thank you to all those who have supported us, and continue to support us. At least two more years to work on insurance and predictive models, fairness, calibration, discrimination, trust, explainability, interpretability, market equilibria, competition, generative models, and so much more… We’ve still got a lot of work to do, and plenty of enthusiasm!

Students

Bayesian Improved Surname Geocoding to predict “Race” in the U.S.

Leave a comment

After Florent last week, this Wednesday, Ana (Patrón Piñerez) will give a talk to conclude her internship in Montréal (supervised with Agathe), on Bayesian Improved Surname Geocoding to predict “Race” in the U.S.

This study focuses on predicting race within the United States, a topic of significant sensitivity due to legal prohibitions against discrimination based on ‘race, color, or previous condition of servitude’ (Civil Rights Act of 1866, 1964). At the same time, due to the rising prevalence of big data systems, insurers are increasingly required to adhere to regulations such as Colorado SB21-169, which mandate demonstrating non-discriminatory practices toward sensitive attributes such as race. In this context, our research explores various methodologies for race prediction. We begin by examining pre-Bayesian approaches that utilize geolocation and surname data, progressing to Bayesian methods that integrate these sources. Specifically, we discuss Bayesian Improved Surname Geocoding (BISG) and its adaptations. Challenges associated with these techniques, such as handling zero counts for minorities and data scarcity, are also addressed. Finally, we propose a novel strategy known as nested dichotomies, rooted in the BISG algorithm. Unlike traditional multiclass prediction, this approach involves sequential binomial predictions structured within a nested framework.