Mining configurable enterprise information systems

Process mining aim at extracting a process model from system logs. These logs have to meet minimum requirements, i.e. each event should refer to a case and a task. Some system logs do not meet these requirements, and therefore it is not possible to use process mining for process optimization or delta analysis. This paper shows an alternative process mining procedure for logs containing only data on the number of times that process steps have been executed (i.e., frequencies). To be able to mine such logs we apply Configurable Event-driven Process Chains (C-EPCs). If a C-EPC is available, we propose a method to mine the process. If only a classical reference model (i.e., an EPC) is available, we propose a method to first derive the C-EPC through mining and then analyze the process. This approach enables us to do process mining in the context of ERP systems such as the SAP solutions. (c) 2005 Elsevier B.V. All rights reserved.