explainable_concept_drift_pm

This is the implementation of the Framework for Explainable Concept Drift Detection for Process Mining Paper https://arxiv.org/pdf/2105.13155.pdf, built ontop of PM4Py https://pm4py.fit.fraunhofer.de/.

Installation

Under Python 3.7, use pip to install the requirements

pip install -r requirements.txt

If you use anaconda just run the commands specified in setup.txt, this will setup a new environment under pyhton 3.7 and install all required packages.

Usage

To run the two examples specified in the paper, run either

python synthetic.py

or

python bpi_2017.py

The output will be written to the corresponding pdfs.

In order to run the bpi_2017.py script, please unzip the log in pm4py/statistics/time_series/experiments/data into BPI2017.xes