| SUMMARY: |
The thesis is placed in an emerging research area from the last decade: process mining. The goal of process mining is to transform data from event logs into information about the actual processes that take place in organizations. The goal of employing process mining techniques on the event logs produced by information systems used in companies, is to understand what actually happens in the organization in order to improve the internal business processes. Generally, process mining methods and techniques provide a control-flow perspective of the process, ignoring the data-flow perspective. The author became aware that recently there is an increased interest shown for the data perspective (e.g. many employees from real companies are asking questions like “if I am in a certain state of a process, what data do I know so far and what data will I need to find out in order to complete the process?”). This motivated us to pursue a mining approach on event logs aimed at extracting a data-flow model. To operationalize this, three process mining algorithms which offer a data-centric perspective of a process are introduced in this thesis. The starting point, similarly to process mining, is represented by event logs. The author introduced the problem context in the first chapters of the thesis, then the author presented our approach and, finally, we validated our claims using various event logs. |
|
