News > 2020 > Maxime Larocque, winner of the Best Master’s Thesis Award

Maxime Larocque, winner of the Best Master’s Thesis Award

May 5, 2020

Maxime Larocque, a graduate of the Master of Business Intelligence (MSc) program, has received the 2019 Best Master’s Thesis Award.

Maxime Laroque’s thesis, Sélection objective de variables à l’aide d’algorithmes génétiques ensachés (Bagged genetic algorithms for objective variable selection), was co-directed by professors Jean-François Plante and Michel Adès (UQAM). The key innovative approach that was used and impressed the jury was founded on thorough, high-quality work.

Laroque, who is now a data scientist for Ubisoft Montréal, states:

“In this era of big data, we explored the fundamental problem of selecting variables—a methodological challenge dating back several decades but for which there is no perfectly satisfactory solution. To do this, we used genetic algorithms, an approach that is not well known in the statistics community.”

Whether for classical statistics or artificial intelligence, variable selection makes it possible to identify the most effective models by choosing the most relevant predictors. This is an essential data science tool for supervised learning when predicting a target variable using structured and tabular data.

A method that can be applied to multiple situations

This master’s thesis provides a methodological contribution. It consists of developing a “genetic” method, which makes it possible to identify the right subsets by imitating the mechanisms of evolution by natural selection, which applies to a vast number of situations.

For example, the target variable can be the prevalence of a rare disease, the market valuation of a commodity, the ability to walk six months after a spinal cord injury, the attrition rate of clients over a 30-day period or the monetization habits of a video-game player.

With the results of this thesis, variable selection based on a parallel universe can now be used automatically and objectively—two characteristics often needed to ensure an analysis is useful and credible.

In addition to submitting an article based on his thesis to the prestigious Journal of the American Statistical Association, Maxime Larocque has successfully applied his research results to the car insurance and video game industries. His method has also been shared at academic and professional conferences (e.g., STATQAM, Ubisoft ML Roundtable).

Congratulations to Maxime Larocque and the 13 other Best Master’s Thesis Award finalists for their excellent work!


William Blais


Charles Cayrat

Organizational Development

Allison Drouet-Chen


Antoine Grenier Ouimet

IT Business Analysis

Salma Hassani Alaoui

IT Business Analysis

Sergej Lackmann

Business Intelligence

Élisabeth Lasnier


Paul Léné

Organizational Development

Thomas Mayer-Chéret


Gabrielle Naime Mourra


Anne-Sophie Prémont-Picard

International Affairs

Elie Saaoud


Mamadou Yamar Thioub

Financial Engineering