User Behavior Mining (UBM)
User Behavior Mining investigates the behavior of software users by means of process mining methods. A detailed observation of real user behavior is essential for software companies, as it can reveal whether the actual behavior corresponds to the intended use. If this is not the case, the efficiency of the executed business processes as well as the usability of the software can be negatively affected.
User behavior mining tracks how users interact and work with the software. For this purpose, existing methods from process mining and machine learning are adapted, further developed, and applied to UBM data. The data basis is provided by so-called UI logs, meaning high-resolution event logs in which the individual interactions between the users and the system are recorded. This way, human behavior can be analyzed to develop assistance components, identify automation potentials, or confirm or refute corresponding theories. User behavior mining thus provides the possibility to collect and analyze empirical data on behavior patterns of software users over a longer period of time and draw new interesting insights.
Applicability of Process Mining in Business Practice
The idea of process mining is to use process execution logs that are automatically collected in many IT systems for an improved management of business processes. It combines process-oriented management and business process modeling with data analysis, data mining, and machine learning methods. However, the process mining methods developed in research are mostly analytical in nature; their practical added value is often not directly apparent to companies. In this respect, it is of central importance to (further) develop these methods regarding the direct generation of entrepreneurial added value and better usability by non-experts. For example, this enables an automated comparison of process flows with textual process descriptions, such as manuals.
This research focus also includes the development of standardized methods for the evaluation of process mining methods, especially process discovery approaches. The goal is to make process discovery approaches more comparable and thus to promote their further development.