EDImine - Mining Inter-organizational Business Processes
By zapletal - Posted on November 9th, 2010
Typical approaches to business process management capture what a company wants to do. In contrast, process mining reveals what a company is actually doing by discovering and analyzing business processes based on event logs. So far, process mining approaches have concentrated on the internal business processes of companies. The project EDImine seeks to extend current process mining approaches in order to apply them to inter-organizational business processes.Choreography modeling and service integration have been well researched in current literature. In contrary, real-world implementations of inter-organizational systems are still (and will be for some time) realized by traditional Electronic Data Interchange (EDI) standards. In traditional EDI standards, the notion of process or choreography is not explicitly specified. Rather, every business document exchange stands for its own.Thus, the first research question is, whether we are able to derive the choreography of an inter-organizational business process based on the EDI messages that are interchanged between companies. The second research question deals with the fact that a set of EDI message exchanges (request for quote, order, invoice, etc.) comprises a lot of redundantly transferred business information carried out in the same business case. In practice, every piece of redundantly transferred information elevates the risk of semantic heterogeneity, since each involved business partner has to support its processing. Moving towards a choreography-based approach for EDI systems, however, presupposes only the communication of the appropriate"delta"of business information in order to handle the next step in the process. Thus, our second research question is whether we can identify the minimum as well as the redundant part of information that is communicated in the course of a discovered EDI process.Finally, the EDImine project seeks not only to mine the interaction sequences between the partners, but also to derive business decisions from the obtained information. This goes beyond current state-of-the-art in process mining, since it combines process information and business-relevant data. Hence, EDImine's third research question is whether we can lift the information gained on the IT level (from the event logs as well as from the messages) to the business level. The goal is to answer questions not only regarding interaction performance, but also with respect to business performance. Combining interaction with business performance leads to new insights into the"value"of business partnerships. By dealing with the three research questions outlined above, EDImine's objective is to support companies in decision making with respect to their inter-organizational relationships.