Improving Data Quality, Model Functionalities and Optimizing User Interfaces in Decision Support Systems

This dissertation contributes to the research on three core elements of decision support systems for managers and consumers: data management, model management and user interface. With respect to data management this dissertation proposes an approach for reducing unobserved product heterogeneity in online transaction data sets. The example of an online auction data set is used to investigate the approach’s ability to improve data quality. In the area of model management this dissertation contributes an approach to elicit consumer product preferences for exponential (beside linear) utility functions aiming at predicting consumers’ utilities and willingness-to-pay for individual products. The question which utility function (linear or exponential) is better suited for predicting product utilities and the willingness to pay is evaluated using a laboratory experiment. Further, in the area of user interfaces this dissertation deals with information visualization. Focusing on coordinate systems, a laboratory experiment is used to investigate which visualization format (two or three dimensional) is better suited for supporting simple vs. complex decision making scenarios and which criteria matter when choosing a visualization format for a particular level of decision making complexity.