Dijana Oreški - MLC 2023

Dijana Oreški

Associate Professor, University of Zagreb, Faculty of Organization and Informatics

Dijana Oreški works as associate professor at the University of Zagreb, Faculty of Organization and Informatics.She has received her PhD in the field of data science, developing new techniques for feature selection in data mining process. She has published over 70 scientific articles in the field of data science. Her research work is focused on the implementation of machine learning algorithms in social science. She has worked on the numerous international projects in the related field.

Social Profiles

All Sessions by Dijana Oreški

  • II DAN

    Subota, 21. maj 2022.

  • 11:00  - 11:30
    Over the last few years we have witnessed a huge number of machine learning algorithms development and their applications in a broad spectrum of domains. The application of these algorithms and the process of developing quality models is complex, iterative and time-consuming. There is a need to automate that process and determine which algorithm is best to use in a specific situation, in a particular domain. Previous research in the domain of classification techniques suggests that dataset characteristics considerably impact the performance of the machine learning algorithms and proves that the choice of the “best” classification algorithm is dependent on the given dataset. In this talk, focus is on the identification of educational data characteristics, their usage in meta-models development and finding patterns between data characteristics and machine learning algorithms performance to achieve step closer to automated machine learning.
    Where
    Velika Terasa, Palić




Share

Dijana Oreški