Identification of Advanced Activity Concepts in SAP for Process Mining

 

Identification of Advanced Activity Concepts in SAP for Process Mining

Student: Kapil Deshmukh

Title: Identification of Advanced Activity Concepts in SAP for Process Mining

Supervisor: Alessandro Berti

1st Examiner: Prof. Dr. Wil van der Aalst

2nd Examiner: Prof. Dr. Ulrik Schroeder

Summary:

Process mining allows users to gain insights into complex business processes by analysing event logs. Object-centric process mining is a new branch of process mining, which facilitates the exploration of interconnected processes, like the ones recorded by SAP systems. This new exploration technique requires Object-centric event logs (OCEL) as inputs, where each event is related to multiple objects. SAP systems are a rich source of raw data for Object-centric process mining. However, extracting OCEL from SAP is challenging due to its complex relational schema and the need for high domain knowledge. This master thesis proposes an approach to simplify OCEL extraction from the SAP systems, by introducing Advanced Activity Concepts. This method combines machine learning with a generic set of rules derived from common patterns in SAP tables, to identify key table attributes to extract events and automatically relate them to various objects in the process. A prototype tool is also developed, implementing the proposed approach, capable of extracting OCEL from the SAP tables with minimal user input. Events extracted using Advanced Activity Concepts have more meaningful descriptions and cover a broader range of activities than existing methods. The performance of the machine learning model is evaluated for accuracy. OCELs are extracted for standard SAP processes using the tool and evaluated for interpretability using the recent developments of leveraging Large Language Models in process mining.

Contact

Name

Alessandro Berti

Software Engineer

Phone

work
+49 241 80 21912

Email

E-Mail