Doctor of Philosophy
School of Computing and Information Technology
As organizations perform their business, they analyze, design and manage a variety of processes represented in models with different scopes and scale of complexity. Specifying these processes requires a certain level of modeling competence. However, this condition does not seem to be balanced with adequate capability of the person(s) who are responsible for the task of defining and modeling an organization or enterprise operation.
On the other hand, an enterprise typically collects various records of all events occur during the operation of their processes. Records, such as the start and end of the tasks in a process instance, state transitions of objects impacted by the process execution, the message exchange during the process execution, etc., are maintained in enterprise repositories as various logs, such as event logs, process logs, effect logs, message logs, etc. Furthermore, the growth rate in the volume of these data generated by enterprise process execution has increased manyfold in just a few years.
On top of these, models often considered as the dashboard view of an enterprise. Models represents an abstraction of the underlying reality of an enterprise. Models also served as the "knowledge driver" through which an enterprise can be managed. Data-driven extraction offers the capability to mine these knowledge drivers from enterprise data and leverage the mined models to establish the set of enterprise data that conforms with the desired behaviour.
This thesis aimed to generate models or knowledge drivers from enterprise data to enable some type of dashboard view of enterprise to provide support for analysts. The rationale for this has been started as the requirement to improve an existing process or to create a new process. It was also mentioned models can also serve as a collection of effectors through which an organization or an enterprise can be managed.
The enterprise data refer to above has been identified as process logs, effect logs, message logs, and invocation logs. The approach in this thesis is to mine these logs to generate process, requirement, and enterprise architecture models, and how goals get fulfilled based on collected operational data.
The above a research question has been formulated as "whether it is possible to derive the knowledge drivers from the enterprise data, which represent the running operation of the enterprise, or in other words, is it possible to use the available data in the enterprise repository to generate the knowledge drivers?".
In Chapter 2, review of literature that can provide the necessary background knowledge to explore the above research question has been presented. Chapter 3 presents how process semantics can be mined. Chapter 4 suggest a way to extract a requirements model. The Chapter 5 presents a way to discover the underlying enterprise architecture and Chapter 6 presents a way to mine how goals get orchestrated. Overall finding have been discussed in Chapter 7 to derive some conclusions.
Santiputri, Metta, Data-driven conceptual modeling: how some knowledge drivers for the enterprise might be mined from enterprise data, Doctor of Philosophy thesis, School of Computing and Information Technology, University of Wollongong, 2017. https://ro.uow.edu.au/theses1/234