Interpretable Knowledge Gain Prediction for Vocational Preparatory E-Learnings

Benjamin Paaßen; Malwina Dywel; Melanie Fleckenstein; Niels Pinkwart

In: Maria Mercedes Rodrigo; Noburu Matsuda; Alexandra I. Cristea; Vania Dimitrova (Hrsg.). Artificial Intelligence in Education. Posters and Late Breaking Results, Workshops and Tutorials, Industry and Innovation Tracks, Practitioners’ and Doctoral Consortium. International Conference on Artificial Intelligence in Education (AIED-2022), July 27-31, Durham, United Kingdom, Pages 132-137, LNCS, Vol. 13356, Springer, Cham, 7/2022.


Vocational further education typically builds upon prior knowledge. For learners who lack this prior knowledge, preparatory e-learnings may help. Therefore, we wish to identify students who would profit from such an e-learning. We consider the example of a math e-learning for the Bachelor Professional of Chemical Production and Management (CCI). To estimate whether the e-learning would help, we employ a predictive model. Developing such a model in a real-world scenario confronted us with a range of challenges, such as small sample sizes, overfitting, or implausible model parameters. We describe how we addressed these challenges such that other practitioners can learn from our case study of employing data mining in vocational training.


German Research Center for Artificial Intelligence
Deutsches Forschungszentrum für Künstliche Intelligenz