Doctoral dissertation of Dr. Simon Staudinger, MSc
In his dissertation, Simon investigated the current and highly important question of how decision makers can assess the reliability of analytical results, such as predictions obtained through the application of machine learning. Machine learning belongs to the probabilistic-statistical branch of artificial intelligence, which means that the results may or may not be accurate. If decision makers could better assess the reliability of individual predictions, this would naturally lead to significantly better decisions, as decisions based on unreliable analytical results can be potentially disastrous.
For those interested: The general approach was originally described in an issue of Springer Nature Computer Science:
Simon Staudinger, Christoph G. Schuetz, Michael Schrefl: A Reference Process for Assessing the Reliability of Predictive Analytics Results. SN Comput. Sci. 5(5): 563 (2024)