Efficient discovery of Declare models from event logs
Name
Taavi Kala
Abstract
The purpose of the following thesis is to improve the performance of a process mining tool called DeclareMiner. The DeclareMiner represents process models in a declarative modelling language called Declare which is widely used in process mining due to its readability and understandability. The current version of the Declare Miner was underperforming in some cases. To improve the performance of the tool, an approach from, which uses replayers, has been integrated into the original algorithm. The result is a new, standalone application whose performance is multiple times better than the existing one. In addition, the new implementation is able to provide different outputs among which a user-friendly and readable report, understandable to people not expert of process mining and Declare.
Graduation Thesis language
English
Graduation Thesis type
Bachelor - Information Technology
Supervisor(s)
Fabrizio Maria Maggi
Defence year
2015