Process mining is an emerging discipline whose aim is to discover, monitor and improve real processes by extracting knowledge from event logs representing actual process executions in a given organizational setting. In this light, it can be applied only if faithful event logs, adhering to accepted standards (such as XES), are available. In many real-world settings, though, such event logs are not explicitly given, but are instead implicitly represented inside legacy information systems of organizations, which are typically managed through rela- tional technology.
In this work, we devise a novel framework that supports domain experts in the extraction of XES event log information from legacy relational databases, and consequently enables the application of standard process mining tools on such data. Differently from previous work, the extraction is driven by a conceptual representation of the domain of interest in terms of an ontology. On the one hand, this ontology is linked to the underlying legacy data leveraging the well- established ontology-based data access (OBDA) paradigm.
On the other hand, our framework allows one to enrich the ontology through user-oriented log extraction annotations, which can be flexibly used to provide different log-oriented views over the data. The annotations are then automatically exploited to understand the legacy data in terms of XES concepts. This, in turn, provides the basis for the process mining algorithms to extract this information either by materializing it explicitly, or by accessing it on-demand. The framework has been implemented in a prototype ProM plug-in that relies on the state-of-the-art OBDA system Ontop.
In this work, we devise a novel framework that supports domain experts in the extraction of XES event log information from legacy relational databases, and consequently enables the application of standard process mining tools on such data. Differently from previous work, the extraction is driven by a conceptual representation of the domain of interest in terms of an ontology. On the one hand, this ontology is linked to the underlying legacy data leveraging the well- established ontology-based data access (OBDA) paradigm.
On the other hand, our framework allows one to enrich the ontology through user-oriented log extraction annotations, which can be flexibly used to provide different log-oriented views over the data. The annotations are then automatically exploited to understand the legacy data in terms of XES concepts. This, in turn, provides the basis for the process mining algorithms to extract this information either by materializing it explicitly, or by accessing it on-demand. The framework has been implemented in a prototype ProM plug-in that relies on the state-of-the-art OBDA system Ontop.