This paper reports on an exploratory study utilizing data mining techniques to predict leadership constructs based on game play data. The learning objective of the game is (1) to become aware of devilish dilemmas during crisis situations, and (2) to understand ones’ leadership style in dealing with these dilemmas. Do player’s act like a People person, as an Administrator, or more like a Figurehead. We evaluate several data mining techniques to predict scoring on these ‘classes’. Our data set consists of 21600 instances. This data was captured over the last 4 years over the course of numerous training sessions for professionals in crisis management organizations in the Netherlands. We found that some algorithms perform significantly better than others in terms of predicting scoring on our test data. Our aim is to develop robust predictive models on the basis of which learning instructions could be given to the trainees during game play to increase their learning journey. However, we conclude that fit for purpose predictive models depend on domain knowledge in the specific field of application.
CITATION STYLE
de Heer, J., & Porskamp, P. (2019). Predictive analytics for leadership assessment. In Advances in Intelligent Systems and Computing (Vol. 783, pp. 516–523). Springer Verlag. https://doi.org/10.1007/978-3-319-94709-9_51
Mendeley helps you to discover research relevant for your work.