The article „Predicting problem-solving success in an office simulation applying N-grams and a random forest to behavioral process data“ explores ways to predict students' success in problem-solving within computer-based simulations early on. The goal is to enable adaptive learning environments that provide tailored support for individual needs.



Using a random forest model, the study analyzed students' behavior during the first 20 minutes of a 55-minute business scenario simulation. This analysis was based on mouse clicks and keyboard inputs. The predictions, especially from the first 10 and 20 minutes, proved to be accurate. These findings highlight how personalized learning support can be provided to less successful students, while adaptive tasks can be designed for more advanced learners using log data.



The article by Sabrina Ludwig, Andreas Rausch, Viola Deutscher, and Jürgen Seifried is available here.