How do I run the Passage package in Eclipse

a_lovely_boy

I imported the Passage package in SVN, but now the JAR package is not the same as the ivy2 jar package that was automatically imported.我想问你我该怎么修改。

hverbeek

Dear,

The Passage package is an older package that has not been kept up-to-date. I've taken a quick look, and it seems that some of the functionality it needs from other packages has been changed in the mean time. That means there is no quick way to get that package working again.

The work in the Passage package has been superseded by the work in the DivideAndConquer package, in which the concept of the passages has been generalized. Perhaps you can start from this package?

Kind regards,

Eric.

a_lovely_boy

Dear, Thank you for your patient reply.There's one more question I'd like to ask.Based on my understanding of your reply,I want to use the current DivideAndConquer package to build Causal Activity Graph from logs and run the Passage algorithm as input to the algorithm. Other plug-ins for project logs, mining models, and merging submodels are all in the current DivideAndConquer package.I would like to ask your opinion, is my idea feasible? Kind regards, wk.

a_lovely_boy

@hverbeek Dear, Thank you for your patient reply.There's one more question I'd like to ask.Based on my understanding of your reply,I want to use the current DivideAndConquer package to build Causal Activity Graph from logs and run the Passage algorithm as input to the algorithm. Other plug-ins for project logs, mining models, and merging submodels are all in the current DivideAndConquer package.I would like to ask your opinion, is my idea feasible? Kind regards, wk.

hverbeek

Dear,

That has all been implemented in the DivideAndConquer package and all packages it depends on. The term 'passages'however is not really being used anymore. But you can build a Causal Activity Graph from a model, then split either an event log or a model (Petri net), then either discover models from the splitted event log or replay the splitted event logs on the splitted model, and afterwards everything is aggregated again in a single event log and a single model.

For the details, I refer to the following paper:

H. M. W. Verbeek, J. Munoz-Gama, and W. M. P. van der Aalst, “Divide and conquer: a tool framework for supporting decomposed discovery in process mining,” The Computer Journal, vol. 60, iss. 11, pp. 1649-1674, 2017.

If you have access to this journal, this paper is available at http://dx.doi.org/10.1093/comjnl/bxx040.

This framework implements the approach as detailed in the following paper:

W.M.P. van der Aalst. Decomposing Petri Nets for Process Mining: A Generic Approach. Distributed and Parallel Databases, 31(4):471-507, 2013.

A preprint of this paper is available at http://www.padsweb.rwth-aachen.de/wvdaalst/publications/p721.pdf.

Kind regards,

Eric.

a_lovely_boy

@hverbeek Dear, Thank you again for your patience.These days I have read the references you provided to me, and I have also run some related plug-ins in ProM.Discover Using Decompostion could not configure the method for finding Causal Activity Graph. So I found another reference for this approach.Discover Matrix,Filter Graph,Create Clusters,Split Event Log,Discover positive Petri Nets, and finally Merge Accepting Petri Nets. But the final net does not remove redundant places.I would like to ask whether the final Accepting Petri Nets can directly call the Reduce Net method.Can I find this method from the Discover Using Decompostion package. Kind regards, wk.

hverbeek

Dear,

On a single Accepting Petri net, there are some plugins that can reduce it (like the "Remove Structural Redundant Places from Accepting Petri Net" plugin from the AcceptingPetriNetRedundantPlacesReductor package). Reductions can either be soundness preserving or language preserving. As far as I know, these plugins do not take a collection of Accepting Petri nets as input.

Kind regards,

Eric.