The missing link between process mining and model-based process analytics is called data-oriented analytical techniques. The course shares data science skills that can be used to improve and analyze processes across a range of industries. Data storage and analysis are not enough. Data scientists must also link data with process analysis. The process mining technique bridges the gap between model-based process analyses (e.g. simulation and other methods for business process management) and data-centric techniques like machine learning or data mining. The goal of process mining is to confront event data, i.e. observed behavior and process models (handmade or automatically discovered). Although this technology is only now available, it can be used to all types of operations (organizations or systems). Examples of applications are: Analyzing hospital treatment, optimizing customer service in multinationals, understanding customers browsing habits on booking sites, analysing baggage handling failures, and improving an X-ray machine's user interface. These applications all share one thing in common: dynamic behavior must be linked to process models. This is what we call "data science in practice".
This course will explain the fundamental analysis techniques used in process mining. The course will teach participants how to discover process models using various algorithms. They can then be used to learn process models automatically from event data. Other process analysis methods that make use of event data will also be covered. The course provides easy-to use software and real data sets as well as practical skills for applying the theory to a wide range of applications.
The course begins with an introduction to the technologies and approaches that make use of event data for decision-making and process (re)design. Next, the course will focus on process mining which is a bridge between business process modeling and data mining. This course has an introductory level and includes various assignments.
This course will cover the main three types of process mining.
1. Discovery is the first form of process mining. The discovery technique uses an event log to create a process model. It does not require any prior information. The Alpha-algorithm, which takes an event log to create a Petri net (a process model) that explains the behaviour recorded.
2. Conformance is the second form of process mining. A process model and an event log are compared. To check whether the reality as it is recorded in the log conforms to the model or vice versa, you can use conformance checking.
3. Enhancement is the third kind of process mining. Enhancement is a third type of process mining. It involves the extension or improvement of a process model by using data about actual processes that has been recorded in an event log. Conformance checking is concerned with the relationship between the model and the reality. This third form of process mining, however, aims to change or extend the a-priori models. Extend a process model that includes performance information to show bottlenecks, for example. The process mining methods can be applied offline or online. Operational support is the latter. One example of this is detection at the time the actual deviation occurs. Time prediction is another example. This is where a case is partially completed and the processing time remaining is calculated based upon historical information from similar cases.
The use of process mining is not just a way to bridge data mining with business process management, but it can also help address the traditional divide between IT and business. Process mining-based evidence-based management of business processes helps create common ground for information system development and business process improvement.
