For the interested reader pointers follow to the scientific papers that describe in great detail how the Alpha miner works. Note that this is extra material and this is not required for the course.
- W.M.P. van der Aalst, A.J.M.M. Weijters, and L. Maruster. Workflow Mining: Which Processes can be Rediscovered? BETA Working Paper Series, WP 74, Eindhoven University of Technology, Eindhoven, 2002.
- W.M.P. van der Aalst, A.J.M.M. Weijters, and L. Maruster. Workflow Mining: Discovering Process Models from Event Logs IEEE Transactions on Knowledge and Data Engineering, 16(9):1128-1142, 2004.
Note that many extensions of the Alpha miner exist:
- Alpha+: A.K.A. de Medeiros, B.F. van Dongen, W.M.P. van der Aalst, and A.J.M.M. Weijters. Process Mining for Ubiquitous Mobile Systems: An Overview and a Concrete Algorithm. In L. Baresi, S. Dustdar, H. Gall, and M. Matera, editors, Ubiquitous Mobile Information and Collaboration Systems (UMICS 2004), pages 156-170, 2004.
- Alpha++: L. Wen, W.M.P. van der Aalst, J. Wang, and J. Sun. Mining process models with non-free-choice constructs. Data Mining and Knowledge Discovery, 15(2):145-180, 2007.
- Alpha#: L. Wen, J. Wang, W.M.P. van der Aalst, B. Huang, and J. Sun. Mining Process Models with Prime Invisible Tasks. Data and Knowledge Engineering, 69(10):999-1021, 2010.
In short, the differences between the four algorithms are that the Alpha miner is of course the first and most basic implementation. The Alpha+ algorithm is an extension on the Alpha miner which can deal with short-loops and self-loops. The Alpha++ algorithm is again an extension, which is able to find additional non-free choice constructs (e.g. can discover more complex patterns). The Alpha# algorithm can also discover several invisible tasks that are not observable but which do influence the control flow.