In heaven as on earth: The performance of students is as good as it is the digraph that describes their behavior

Detecting groups struggling to overcome their laboratory assignments is paramount for the teacher to assist them, the earliest, the better. Discovering the behavior of students with process mining techniques has proven to be very useful, but commonly associated to students performance indicators. This study proves that student‘s performance indicators might be very helpful, but as much as it is their behavior, strictly from a topological point of view. In this case, using process mining techniques to extract their behavior from a virtual laboratory and using supervised machine learning techniques to identify "bad topological graphs" also provides enough evidence to foresee these groups at risk