zihangs / miningpnmls

miningPNMLS

This repo provides a tool to iterate through a given directory and checks all sub-directories to find all event logs with .xes extensions. And then, it will mine petri-nets (with .pnml extensions) from the event logs using a specific miner with a set of configurable parameters. The details of miners refers to here.

Instructions

1. Prepare the directory which contains the event logs you want mine process models from (the target directory). You can put this directory anywhere, but for simpleness, I just put my target directory within this repo called gene_data/.

2. Select a miner, the miner.jar has wrapped up with 3 miners (inductive miner -IM, directly flow miner -DFM, transition system miner -TSM). Just select one (-IM, -DFM or -TSM).

3. Specify the parameters for each miner:

   -IM:

   • Noise threshold (a float), set 0.8 as default.

   -DFM:

   • Noise threshold (a double), set 0.8 as default.

   -TSM:

   • Number limit of states (a int), set -1 as default (means no limit). But you can set to any positive integer as a limit.
   • The top percentage of event name (a int), set 100 as default, means including all.
   • The top percentage of label (a int), set 100 as default, means including all.

Examples

Target directory ./gene_data/ using -IM with threshold 1.0.

#java -jar mine_all_pnmls.jar <miner> <dir_path> <theshold>
java -jar mine_all_pnmls.jar -IM ./gene_data/ 1.0

Target directory ./gene_data/ using -DFM with threshold 0.8.

#java -jar mine_all_pnmls.jar <miner> <dir_path> <theshold>
java -jar mine_all_pnmls.jar -DFM ./gene_data/ 0.8

Target directory ./gene_data/ using -TSM with no states limit, top 100% event names and top 100% labels.

#java -jar mine_all_pnmls.jar <miner> <dir_path> <limit> <top_percent_event_name> <top_percent_label>
java -jar mine_all_pnmls.jar -TSM ./gene_data/ -1 100 100

Outputs

The newly mined petri-net models (.pnml files) will stored in the same place with the original event logs (.xes files). And the file name will be <original_name.xes.pnml>. In my case, the output .pnml files are stored in ./gene_data/.

Notice: this tool was developed with java 8 runtime environment, so better to use java 8 to avoid issues.