Back
 

15 January 2020

DRiM Game 2019: the data science challenge for students

Last December, about forty students took up the challenge offer by DRiM Game, an initiative supported by RCI Bank and Services as a partner. This challenge gave masters’ degree students the opportunity to explore new methods of default probabilities, a key topic for the group.

 

A DATA SCIENCE CHALLENGE ON REAL DATA

For the second time in a row, Within this framework, RCI Bank and Services participated in a university challenge called "DRIM Game" on December 10th and 11th, in partnership with Deloitte and the master ESA from Université d’Orléans,.

The 2019 edition brought together some forty students from four master's programs (Master ESA Université d'Orléans - Aix-Marseille School of Economics / Centrale Marseille - Master Économétrie-Statistique, Université de Paris 1 - Master d'Econométrie-Statistique, IAE Nantes)

Since October 2019, students were assigned a general theme: predicting default probability for a range of periods. At this stage, RCI Bank and Services provided students with a sample of real data so that they could get more familiar with the database and reflect on the methodologies they were going to use.

On Tuesday December 10th, the students gathered at Tour Majunga, in Deloitte's offices at La Défense. They received the complete and definitive database and had until Wednesday afternoon to work on the subject.

On Wednesday afternoon, the eight teams presented their work on modeling default probability before the jury made up of credit risk experts: members of the analytics departments of RCI Bank and Services, Deloitte and SAS, partners as well.

Each team had 15 minutes to present its work, first with a classic statistical modeling methodology, then with a machine learning method.

After the jury's careful consideration, prizes were awarded to the following teams:

  • 1st: group 2 (Master ESA, Université d'Orléans)
  • 2nd: group 1 (Master ESA, Université d'Orléans)
  • 3rd: group 6 (Aix-Marseille School of Economics / Centrale Marseille)

CREDIT RISK MODELS, A KEY TOPIC FOR RCI BANK AND SERVICES

Credit risk is the risk that a borrow will not repay all or part of their loan by the deadlines outlined in their contract. The management of credit risk is part of RCI Bank and Services' core business as it determines the profitability of transactions carried out.

Credit risk models are used to manage risk and optimize the approval process and coverage, depending on default probability. They also serve to calculate impairment while respecting IFRS9 standards, and give the company the opportunity to be approved by the European Central Bank in order to calculate its own funds requirements using an internal approach.

This challenge allowed us to explore alternative and innovative methods in order to meet the challenges linked to credit risk, in a framework involving the different schools and universities specialized in the subject.

Samy Taouri, Head of Internal Credit Risk Models Department