A domain-specific language for supporting event log extraction from ERP systems

• Published in: 2018 7th International Conference on Computers Communications and Control (ICCCC) pp. 12 - 16
• Main Authors: Simovic, Ana Pajic, Babarogic, Sladan, Pantelic, Ognjen
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01-05-2018
• Subjects: artifact-centric modeling; Business; business artifacts; Convergence; Data mining; domain specific language; Domain specific languages; ERP system; event log extraction; Information systems; Xenon
• DOI: 10.1109/ICCCC.2018.8390430

Process mining techniques provide capabilities for discovering the real business process flows from data, and compare expected and actual behaviors. Actual behaviors, in many cases, are obtained from Enterprise Resource Planning (ERP) systems and other enterprise information systems transaction logs. These transaction logs provide valuable insight into the companies' business processes. They traditionally hold a large amount of data in a set of conceptual documents related to each other through one-to-many and many-to-many relations, where information changes occur in transactions. Underlying data model gives rise to complex interactions between multiple data objects without a clear notion of a unique case identifier in an isolated process. However, enterprise process mining techniques can be applied only to event logs containing event data related to one notion of process instances. Within ERP systems, such event logs are not explicitly given and substantial domain knowledge is required to select the right data from multiple tables in relational databases. In order to respond to this need, in this paper we present an abstract syntax of domain-specific language (DSL) for facilitating the extraction of an appropriate dataset from ERP systems by domain experts, and its conversion into event log based on XES IEEE standard. It is developed specifically to describe behavior over complex data from ERP systems in terms of multiple interacting artifacts. The goal is to align the data and process perspectives, supporting extraction of complex ambiguous cases, affected by data convergence and data divergence problems. The basic concepts of the language as well as principles are discussed in depth in this paper.