22.07.2022

QDS Master student Kento Okuyama and his team won first Datathon of the International Meeting of the Psychometric Society 2022

At the Datathon, teams of 2-4 worked together from June 20th, 2022, to July 1st, 2022, to solve a common data-oriented problem

Master student, Kento Okuyama, from the QDS master program (of the Methods Center) took part in the IMPS Student Datathon which happened during the International Meeting of the Psychometrics Society (IMPS) 2022 in Bologna, Italy.

At the Datathon, teams of 2-4 worked together from June 20, 2022 to July 1, 2022 to solve a common data-oriented problem provided by Math Garden. The competition was sponsored by the University of Amsterdam and the winner would receive a cash prize of $1,000.

From given data of an online learning platform an IRT model was estimated. The item difficulties were originally computed based on the ELO algorithm that considers the matching history of each item and person.

The teams, however, did not get all the item difficulties but had to estimate some missing item difficulties while only having the multiplication items as text (e.g., 6 x 3, indicating the perceived difficulty of the item) with some other information (e.g., response time per item) as well as the estimated item difficulties of the other items.

Therefore, the teams had to come up with machine learning methods to maximize the prediction of the missing item difficulties.

Based on the handed-in result, Kento Okuyama and his team members (Rosa Fabbricatore from University of Naples Federico II and Matteo Orsoni from University of Bologna) were selected as one of the top teams to deliver a final presentation on July 15th the IMPS in Bologna. Consequently, Kento Okuyama and his team were selected as the winning team by the commitee. Congratulations!