The notebooks in this repository are part of the assignment in the course Advanced Process Mining and intended as a starter for building your own prediction models for predictive process monitoring. They can be used as:
- Cloud notebooks via MyBinder
- Local stand-alone notebooks
- Local Dockerized notebooks
refer to the Installations & Usage section below for usage instructions.
You may also refer to the PM4Py documentation on Machine Learning for further options or an alternative to this implementation: https://pm4py.fit.fraunhofer.de/static/assets/api/2.7.8/api.html#machine-learning-pm4py-ml
The collection of notebooks is a living document and subject to change.
Click on the launch binder links for either the R or the Python notebook. You could also use the Google Colab service; however, you may need to prepare the Google Colab environment to have the respective packages installed (see standalone instructions).
Build a Docker image with the provided Dockerfile:
docker build -t fmannhardt/starter-predictive-process-monitoring .
And start the Docker container running Jupyter on localhost:8888:
docker run --rm -ti -e JUPYTER_TOKEN=processmining -p 8888:8888 fmannhardt/starter-predictive-process-monitoring
or use the Jupyter Lab interface:
docker run --rm -ti -e JUPYTER_TOKEN=processmining -p 8888:8888 fmannhardt/starter-predictive-process-monitoring sh -c "jupyter lab --ip 0.0.0.0 --no-browser"
You should be able to run the Jupyter notebooks directly in a Jupyter environment using:
jupyter lab
Please make sure to have installed the following requirements:
Python
pip install -r requirements.txt
Make sure to install GraphViz for the visualization. On Windows with Chocolately this should work:
choco install graphviz
Consult the PM4Py documentation for further details for other environments.