Process Mining for Adpative Process-driven User Guidance (Master)

In particular Business Analytics solutions are designed to solve complex tasks with highly expertise Visual Analytics interfaces. The tools itself are almost quite effective, however, the learning and usage of these tools have a strange learning curve. This leads to confused and overstrained users, in particular those users at beginner level, who have many problems in using the software correctly and efficiently.

Hints and guidance approaches can help to make such kind of software better usable, but these approaches are statically designed and therefore require heavy efforts to prepare and train them. Process mining algorithms on the other hand aiming on automatically generation of processes on the basis of events.

The goal of this master thesis is to a solution that automatically learns processes and tasks on the basis of system events, in particular user interaction events. In a second step the solution should be able to recognize a process/task and provide automatically the best tools (i.e. visualizations) to solve it.


Dipl.-Inf. Dirk Burkhardt