A Framework for Extracting and Encoding Features from Object-Centric Event Data

Jan Niklas Adams, Gyunam Park, Sergej Levich, Daniel Schuster, Wil van der Aalst

If you find the provided code or the corresponding paper useful, please consider citing the paper:

@inproceedings{AdamsFeaturesOCEL10.1007/978-3-031-20984-0_3,
  author    = {Jan Niklas Adams and
               Gyunam Park and
               Sergej Levich and
               Daniel Schuster and
               Wil M. P. van der Aalst},
  editor    = {Javier Troya and
               Brahim Medjahed and
               Lina Yao and
               Pablo Fern{\'a}ndez and
               Antonio Ruiz-Cort{\'e}s},
  title     = {A Framework for Extracting and Encoding Features from Object-Centric Event Data},
  booktitle = {Service-Oriented Computing - 20th International Conference, {ICSOC}
               2022, Sevilla, Spain, November 29- December 2, 2022, Proceedings},
  series    = {Lecture Notes in Computer Science},
  volume    = {13740},
  pages     = {36--53},
  publisher = {Springer},
  year      = {2022},
  url       = {https://doi.org/10.1007/978-3-031-20984-0_3},
  doi       = {10.1007/978-3-031-20984-0_3},
}

The paper is available at: Paper

To get an up-to-date version of the python library used please visit ocpa.readthedocs.io

---

Instructions

Please first use anaconda prompt to create the environment.

conda env create --file environment.yml

then unzip the event log in example_logs/mdl/ and put it in the same directory.

Activate the environment in anaconda prompt

conda activate icsoc

Go into the repository directory and run

python experiments.py

This will run all the experiments and reproduce the figures. Slight deviations might be due to differently initialized weights in the neural networks.

Note: To use the DGL library, you additionally have to place a config file in

~\.dgl\config.json

containing only the line {"backend":"tensorflow"}

There might be an error if the environment was installed with a wrong version of a package. https://stackoverflow.com/questions/72441758/typeerror-descriptors-cannot-not-be-created-directly

run pip install protobuf==3.20.* in that case.