Dataset for Algorithm Evaluation

lubken

Hey

I am doing at the moment my Master Thesis in the context of Process Mining Algorithm Evaluation. I need for this two datasets to evaluate on. I have one from a company I cooperate with. About the other one I would like to ask if some Datasets are public? Or could be provided to me confidentially?  Like the datasets presented in the Genetic Mining paper ( Bezwaar BezwaarWOZ Afschriften Bouwvergunning ).

Alternative would be creating a dataset artifically (http://www.processmining.it/sw/plg) But I would prefer real life logs.

Thank you for your help.

Best regards

Lukas

JBuijs

Hi,

For the Business Process Intelligence workshop at the BPM'11 conference there is a challenge published. This includes a large, real life, event log.
You might want to check out http://www.win.tue.nl/bpi2011/doku.php?id=challenge and/or http://data.3tu.nl/repository/uuid:d9769f3d-0ab0-4fb8-803b-0d1120ffcf54

I think you can use this event log for you testing purposes.

lubken

Thank you, that looks great.



Is this log used in some article yet? I read the information in the link
but would be interested in deeper analysis of other papers.



Thanks

JBuijs

Hi Lubken,

Well, the idea is actually that a lot of people start to analyse this log and prepare a report in a contest setting. The one that can create the most creative, best, interesting, ... report wins.
So, no, there are no papers that use this dataset, also because this dataset has been released only recently.

There is a master student in our group working on, and almost finished with, creating a dataset to test process discovery algorithms. I will ask him if his dataset is ready.