Skip to content

Stochastic Process Mining

Study level

PhD

Master of Philosophy

Honours

Vacation research experience scheme

Faculty/Lead unit

Science and Engineering Faculty

School of Information Systems

Topic status

We're looking for students to study this topic.

Supervisors

Dr Sander Leemans
Position
Lecturer in Business Process Management
Division / Faculty
Science and Engineering Faculty

Overview

Process mining aims to derive information from historical behaviour of processes in organisations which has been recorded in event logs.

Business analysts use process mining software to visualise logs and derive information and insights for managers. Ultimately, this information is used to improve processes to, for instance, optimise costs, time and/or the environment. Process mining is an exciting field with lots of opportunity for research and with many successful commercial solutions being offered.

Business process models describe what can happen in a process. That is, the order and choices between the steps (activities) that have to be executed to traverse the process and to handle a case such as an insurance claim/process an order/admit a student/etc. Business processes are described by formalisms such as Business Process Model and Notation (BPMN), Petri nets and process trees.

In some cases, the information contained in process models is not sufficient. For instance, if you want to predict the remaining time of a particular case that is running, information of the likelihood of future paths through the model is necessary. Furthermore, to improve processes, it might be worthwhile to focus on the most occuring parts of the process, rather than focusing on exceptional cases.

To solve this, stochastic process models not only express which cases are supported by a model, but also how likely each case is to occur. For instance, different types of stochastic Petri nets have been proposed, in which every transition is given a "weight" and the "heavier" a transition, the more likely that transition is to be fired before other competing transitions.

Several types of stochastic process models have been proposed and a discovery technique, a technique that takes an event log and creates a stochastic process model automatically, has been published.

Research activities

In this project, we aim to research several aspects of stochastic models as well as design new techniques for:

  • How can we discover stochastic process models from event logs automatically
  • How to assess the quality of stochastic process models with respect to event logs or one another (conformance checking)
  • How to best use stochastic process models in practice: how to get the most value for industry partners (how-to manuals, methodologies, user-studies, visualisations, explanations, etc.)
  • What are the limitations of current types of stochastic process models and could we address them.

You will have the option of choosing one or more of these fields.

Outcomes

All of the mentioned research activities are on the edge of current scientific knowledge.

Therefore, we expect to publish the performed research in top conferences and journal articles.

Skills and experience

Depending on the specific topic chosen, we require the following skills:

  • basic programming skills to implement prototypes of designed techniques
  • basic understanding of process models, event logs, process mining and/or business process management
  • basic understanding of mathematics.

Scholarships

You may be able to apply for a research scholarship in our annual scholarship round.

Annual scholarship round

Keywords

Contact

Contact the supervisor for more information.