������

The doctoral thesis of Taeyoung Kee, M.Sc., "Essays on decision analytics: New approaches for modeling preferences, uncertainties, and data" will be publicly examined at the Aalto University School of Business on Thursday, May 21, 2026.

https://aalto.zoom.us/j/65799621062

Organizations across all sectors — from governments planning climate action to companies managing R&D portfolios — routinely face complex decisions about which projects, investments, or actions to pursue. These decisions must often balance multiple, potentially conflicting objectives under limited resources and significant uncertainty. Adding to the challenge, the items under consideration frequently interact with one another: selecting two projects together may create synergies that exceed their individual benefits, or conversely, lead to cannibalization effects that diminish overall value. At the same time, decisions are typically made on the basis of uncertain performance estimates, which can lead to systematic disappointment when actual outcomes fall short of expectations. Traditional decision support models often oversimplify these complexities, while the data sources they rely on may be rigid and outdated. The growing availability of diverse data sources and advances in computational methods present new opportunities to address these challenges.

This dissertation contributes to developing decision support methodologies by presenting novel frameworks in four interrelated essays. The research proposes new approaches to account for project interactions in portfolio selection, manage the impact of estimation uncertainties on decision quality, and leverage alternative data sources such as financial news articles. The overall focus is on building models that are theoretically sound yet practically feasible, balancing model accuracy with the cognitive burden placed on decision-makers.

The contributions of the dissertation are two-fold: First, it offers theoretical foundations and practical tools for tackling complex decision problems — from axiomatic frameworks for preference modeling to optimization models and Bayesian methods for handling uncertainty. Second, it demonstrates how modern computational techniques, including neural network-based text mining, can enhance decision support in a world increasingly rich in data. The research results provide both readily applicable tools and significant contributions to the foundations of decision science.

Opponent: Professor Matteo Brunelli,University of Trento
Custos (Chairperson): Professor Eeva Vilkkumaa, Aalto University School of Business

Contact information: taeyoung.kee@aalto.fi

Thesis available for public display 7 days prior to the defence at: will be informed later

Doctoral theses in the School of Business: