Πειραματική αποτίμηση αλγορίθμων ανάλυσης διεργασιών
View/ Open
Keywords
PLG2 ; Prom ; Αλγόριθμοι ; Μοντέλα διεργασιών ; Εξόρυξη διεργασιών ; Μετρικές ποιότηταςAbstract
The purpose of this thesis is to apply, study and compare four different process analysis algorithms (process mining). The algorithms to be studied and implemented are: Alpha Miner, Inductive Miner, Evolutionary Tree Miner and Heuristic Miner. The purpose of this work is to compare the above algorithms with emphasis on the quality of the outputs. Process analysis algorithms will be used to record quality metrics such as: Simplicity, Generalization, Precision and Fitness.
In order to test the algorithms, it has been used artificial data which represents process models. The process models will be produced using PLG2 software and the study of process analysis algorithms using Prom software.
During the experiments, various customizations were made to the models. The relevant parameters will be the duration of the model activities, the iterations of the execution model, the variations in the amount of model noise, as well as the changes in the relationships between the model processes.