This is one of the bootcamp sessions that has been opened up for a broader audience.
Process Mining is a relatively recent research area which has been gaining rapid traction in the enterprise, aiming to bridge the gap between traditional data mining techniques (supervised and unsupervised learning) and process-based approaches (simulation, business process management). Contrary to traditional machine learning, which commonly starts from flat, tabular data sources, process mining deals with data describing a flow or sequences of activities or events.
Examples of applications include:
– Discovering where customers cancel their order as they browse your website
– Understand the browsing behavior of costumers before booking
– Analyzing the care flow of patients in a hospital as they move from department to department
– Improving the bottlenecks in a helpdesk
– Tracking the flow of work through your team’s social network
– Verify whether organizational compliance rules are adhered to in a credit assessment process
In short: process mining is all about events and flows. This workshop will guide you through the key fundamentals of process mining: participants will learn the difference between process discovery, conformance checking and enhancement, and will learn through hands-on exercises how to get started right away using real-life event data. Moreover, a number of real-life case studies will be presented to give you a head start towards applying process mining in your own application domain.
What you will learn
- What is BPM, data mining, and how is process mining positioned between them
- The difference between process mining and traditional business process intelligence
- What is an event log, how should it look like, what should it contain
- How to work with the XES file format
- What is process discovery, what are the most common algorithms and how do they work
- What is conformance checking, what is meant by replaying an event log or log alignments
- How can process mining be used to make recommendations
- How to predict the remaining running time of a process or the best next action
- How to work with tools such as ProM and Disco