PhD Defense: Kseniia Kurishchenko

In order to obtain the PhD degree, Kseniia Kurishchenko has submitted her thesis entitled:

Novel Mathematical Optimization Models for Explainable and Fair Machine Learning

Nowadays economists widely use Machine Learning approaches in empirical work. Unfortunately, state-of-the-art Machine Learning models are seen as black boxes. In this dissertation, I propose novel Mathematical Optimization models to trade off accuracy and transparency in Cluster Analysis, Supervised Classification, and Treatment Allocation. The thesis contains six chapters where four of which correspond to papers. In my first two papers, I focus on Cluster Analysis. I develop Mathematical Optimization models and numerical solution approaches to find accurate and distinctive explanations for clusters, by means of prototypes and rules. In the third paper, I improve additive tree models for Supervised Classification, such as Random Forests or XGBoost, making them sparser and fairer, while ensuring they are accurate. I propose a Mathematical Optimization model that trades off these three objectives, while scaling well in the number of observations. In the fourth paper, I investigate the Treatment Allocation problem, where one has to decide which individuals will receive treatment and which not. I introduce a Mathematical Optimization model to have accurate heterogeneous treatment effect predictions and a good level of fairness, which will be the basis for the treatment allocation in forthcoming individuals.

The thesis will be available from research.cbs.dk

Primary Supervisor:
Professor Dolores Romero Morales
Department of Economics
Copenhagen Business School
 

Secondary Supervisors:
Professor Ralf Andreas Wilke
Department of Economics
Copenhagen Business School

Professor Emilio Carrizosa
Statistics and Operations Research
University of Seville
 

Assessment Committee:
Professor Lars Peter Østerdal (Chair)
Department of Economics
Copenhagen Business School

Professor Laura Palagi
Department of Computer, Control and Management Engineering
Sapienza University of Rome

Professor Bart Baesens
Information Systems Engineering Research Group (LIRIS)
KU Leuven

Date: 31 May 2024
Time: 10:00-12:00
Online: Teams

Location: Porcelænshaven
Room: PH16A 2.80
Reception: ECON's kitchen on 3rd floor

*The CBS PhD School will host a reception, which will take place immediately after the defence.