This repository contains the source code, the end-user plugins and the evaluation result files for incremental conformance checking, as proposed in the paper "Sampling and Approximation Techniques for efficient Process Conformance Checking".
The approaches provided include:
- Three conformance measures: Fitness, deviating activities and resource attributes related to non-conformant behaviour (only available from code)
- Event-log sampling with statistical completeness guarantees
- Methods for the approximation of said conformance measures
- Quality retainment approaches to further stabilize the sample-based conformance result. (only available from code)
The methods are implemented as end-user plugins for the Process Mining Toolkit. The repository contains all configuration files for the local development and execution of ProM plugins using the Eclipse IDE.
- Clone this repo (for help see this tutorial).
- Install the Eclipse IDE and import the IncrementalPM directory as a new project.
- In Eclipse, install Apache IvyDE, and resolve all dependencies of the imported project.
- You are done - you may run your local copy of ProM using either "ProM Package Manager (IncrementalPM).launch" for the package Manager or "ProM with UITopia (IncrementalPM).launch" for ProM itself.
Note, that you do not need to install the standalone version of ProM, as the project already includes a local copy of ProM. For further instructions on the usage of ProM itself, or the deployment using other IDE's than Eclipse, please see the ProM Getting started Page or consider contacting the ProM Forum.
In the project, two plugins are provided:
- "Check sample-based Conformance using Incremental Conformance Checking" - conducts a run of the sample-based conformance checking algorithm
- "Evaluate Incremental Conformance Checker" - conducts the controlled experiments used for evaluation of the implemented approaches. The set of result files used in the paper, as well as the scripts for the creation of the plots are provided in the directory "evaluation_results".
The result files, figures and scripts plotting those figures, used for the evaluation in the paper, are located in the directory "Evaluation". In the directory the top-level python script "plot.py" generates a set of plots based on the .csv-files located in the directory "csv", which have been generated using the aforementioned evaluation-plugin. For the sake of clarity, the outputted plots are arranged into directories based on corresponding data set.
This work has received funding from the Deutsche Forschungsgemeinschaft DFG, grant number 421921612, and the Alexander von Humboldt Foundation.
We provide our code, under the MIT license.