Demo Process Discovery - pm4knime/pm4knime-document GitHub Wiki
In this part, the demand to discover a process model is shown at first. Next, the concrete workflow is provided as one sample to solve a similar problem. https://github.com/pm4knime/pm4knime-document/blob/master/images/demo-simple-discover.PNG
One example workflow with PM4KNIME is shown below:
https://github.com/pm4knime/pm4knime-document/blob/master/images/demo-simple-discover-wf.PNG
The following steps will guide you to accomplish the task and discover a Petri net from an event log.
- Download the dataset. Extract the event log
running-example.xes
from the downloaded dataset. - Create a workflow in KNIME called "PM4KNIME-Release-01-WF-Discovery".
- Import the event log running-example.xes
- Drag the node Event Log Reader from the category ProcessMining->IO or by searching the related name, like "Event" in Node Repository.
* Configure the **Event Log Reader** with the path to file running-example.xes and click OK to close the dialog.
* Execute the node by right-clicking the node and choosing **Execute** option.
* Inspect the event log by choosing the Output port, which is always located as the last option by right-clicking.
Click the graph to show different perspectives on the event log.
- Drag the node Inductive Miner into the KNIME and connect it with the output port of the event log.
- Configure the settings with the following values.
Select Inductive Miner Type: Inductive Miner - Infrequent Write the Noise Threshold: 0.2 Select Event Classifier: Event Name
- Click OK to close the configuration dialog.
- Execute the nodes.
- Check the generated Petri net in output port. An accepting Petri net is given as the result.
- Export the generate Petri net with node Petrinet Writer.
-
Connect the input port of Petrinet Writer with the output port of node Inductive Miner.
-
Choose the file path, like
\Generated\running-im-pn.pnml
-
Execute the node to save the Petri net.
-
Following those steps, hopefully, you get more familiar with task of process discovery. More miners are found in the category "Process Mining->Discovery". The node and option settings descriptions are listed on the right site if you click the node.