Web interface for machine learning as decision making support

Abstract: The thesis attempts to explore possibilities of implementing a user interface visualization to a machine learning algorithm. As the algorithm in the thesis does not take soft values into account, the visualization is meant to highlight its’ drawbacks and aware the end user of how the results were generated. Therefore the graphical representation is aimed to fortify the algorithm as decision making support rather than decision making. Machine learning algorithms differ thus making a general consensus towards a graphical representation non-existent. In order to emphasize the user interface aspects, the thesis writers gathered a selection of closely related phenomenon and research articles both within the fields of design process and machine learning.