From June 22 to 26, 2026, teaching assistant Andrej Ševa participated in the "Machine Learning for Time Series" summer school, organized by Radboud University in Nijmegen, the Netherlands.
Radboud University, founded in 1923, is among the leading general universities in the Netherlands and is recognized for the quality of its teaching and its research capacities across multiple scientific fields. As part of its offerings, it also organizes the Radboud Summer School, one of the largest summer schools in the Netherlands, with over 50 intensive academic courses at the Bachelor's, Master's, and PhD levels and more than 600 participants from 75 countries. This tradition makes Nijmegen a significant meeting point for researchers from various parts of the world.
The course program in which assistant Ševa participated brought together researchers and PhD candidates from several European institutions with the aim of deepening knowledge in the application of machine learning to time series. Through hands-on work, participants covered topics including classical benchmark models such as ARIMA and naive models, ensemble algorithms such as XGBoost and Random Forest, anomaly detection techniques (e.g., Isolation Forest), and methods for identifying structural change points in data (change-point detection, Bayesian Online Change Point Detection), with all work carried out on real-world datasets.
Participation in such a renowned program also represents an opportunity to establish and strengthen academic contacts with colleagues from abroad, opening space for future cooperation between the Faculty of Economics, University of Banja Luka, and Radboud University and other partner institutions, through the exchange of experience, joint research initiatives, and potential future projects in the field of applied econometrics and statistics.
Gallery
