Towards a framework for the agile mining of business processes

In order to support business processes effectively, their implementation by a process management systems (PMS) must be as close to the real world's processes as possible. Generally, it is not sufficient to analyze and model a business process only once, and then to handle respective business cases according to the defined model for a long period of time. Instead, process implementations must be quickly adaptable to changing needs. A PMS should enable process instance changes and provide facilities for analyzing these instance-specific changes in order to derive optimized process models. In this paper we introduce a framework for the agile mining of business processes which supports the whole process life cycle in an integrated way. Our framework is based on process mining techniques, adaptive process management, and conversational case-based reasoning. On the one hand, it allows annotating execution and change logs with semantical information to gather information about the reasons for ad-hoc deviations, which can then be analyzed by the process engineer (with support from the PMS). On the other hand, it enables the process engineer to adapt process models based on the outcome of these analyses and to migrate related process instances to the new model.