Title: Creating a framework for comparing and selecting process discovery algorithms
Speaker: Demetris Louca
Abstract: Process discovery is the part of Process mining that creates a model to describe the behaviour of an event-log. Whether it is for training future doctors or optimising business processes, finding the underlying model is essential in gaining an understanding of the occurring processes.

There have been multiple algorithms proposed, such as the alpha-miner, heuristic miner, and others. However, there has not been enough work comparing the algorithms or linking their performance to the properties of the data.

This talk aims to outline the main algorithms in process discovery. Then, the overall performance of the algorithms will be compared experimentally by using a varied set of artificial data. But the overall performance of the generated models is linked to the data used, thus the talk aims to associate some of the descriptive statistics of the data-sets to the performance of algorithms.