Disco and ProM

Gadi

Hi,
I recently finished the course "Process Mining : Data Science in Action" and would like to have a short demo using ProM.
In this demo, I am uploading the CSV file into ProM and would like to get the results like in Disco, but having problems.
The results I am looking are in the files "Disco - Performance" and "Disco - Frequency".

Can anyone details the steps I need to take in order to get the same results in ProM?

Some notes : 
1. I am aware of the fact the Disco is using Fuzzy model + Transition system, still I can't replicate the same results using ProM
2. It is fine not to get the same overview (in visualization, pictures) , as long as I am getting the same results and data

I am attaching  
1. CSV file for uploading
2. The output from Disco, after loading the CSV file, which I would like to get in ProM.

Thanks,
Gadi.

JBuijs

Dear Gadi,

My recommended approach would be to discover a process model (Petri net to be more specific), for instance using the inductive miner.
In a next step, replay the event log for conformance/performance projection, which will show you the latter image.
This was also discussed in the MOOC.

You might also try the inductive visual miner plug-in in ProM.

Hope this helps and good luck with your demo!

sjjleemans

Hi Gadi,

Disco indeed uses ideas of the Fuzzy Miner. However, Disco has been developed further, but how it precisely works is a trade secret of Fluxicon and unknown to the ProM team. Hence, I cannot tell you how to replicate the results of Disco in the Fuzzy Miner.

However, what can be done is to translate the Disco image into a process model. (I made an attempt: see the attached zipped .ptml; as Joos said: it is also possible to use the Inductive Miner or the Inductive visual Miner). Performance can be measured by unzipping this file, importing it into ProM, importing your event log into ProM, and running the plug-in 'visualise deviations on process tree'. For instance, choose to Show 'paths and sojourn times'.

Sander