Creating Process Mining Model

From QPR ProcessAnalyzer Wiki
Jump to: navigation, search

This page describes how a new process mining model can be created in QPR ProcessAnalyzer. To create a model that uses datatables, the following steps need to be done:

  1. Create a datatable for events (eventlog)
  2. (optional) Create a datatable for case attributes.
  3. Create a model that use the datatables.

Model can also fetch the eventlog data using a loading script or an ODBC query. Here are instructions how to create a model using datatables:

  1. Go to the Workspace (by clicking in navigation menu or in the Home screen).
  2. Create a new project for the model by clicking the parent project and clicking NEW -> Project. (Alternatively you can use an existing project.)
  3. Make sure the target project is selected, click NEW -> Datatable.
  4. Select the created datatable and Import.
  5. Select a CSV file, adjust settings for the file and click Start import. (If you don't have any data available, you can try out the sample files).
  6. For the case attributes data, create also a new datatable and import data to it. (This is optional, as case attributes are not mandatory to create a model.)
  7. Open the models list in the header, and find the target project. Hover the project and from the menu (dots icon) select Create model.
  8. Model Properties dialog opens. Define a name for the model.
  9. Open the Cases datasource tab, select the cases datatable, and to the Case ID field define the column that contains the case id's.
  10. Open the Events datasource tab, select the events datatable, and define the columns that contains the case id's, event type and event time.
  11. Click the Save button to close the dialog.
  12. The model has now been created! You can start analysis by going to the Process Discovery and select the created model from in models list in the header.