Integration of machine learning and interactive visualizations for cognition friendly decision making

This thesis investigates how to assist an expert in decision making based on the exploration of classification results. Specifically, this thesis focuses on how to visualize and explore classification results for different data representations in a cognition friendly way. We break the analysis of this research question down into four chapters, starting by studying how to visualize the decision boundary of multi-dimensional classification models in 2D, followed by the application of the proposed decision boundary visualization technique for decision making in risk assessment and finally by adding the notion of prototypes to our studies and evaluating their relevance in decision making and for amplifying cognition.