This project contains the source code for a novel Prescriptive Business Process Monitoring technique (PrBPM) technique for recommending next best actions. Note: The project's source code based on the source code of the paper "Predictive business process monitoring with LSTM neural networks" from Tax et al. (2017).

If you use the code or fragments of it, please cite our paper:

@inprocessings{weinzierl2020nba,
    title={Precriptive Business Process Monitoring for Recommending Next Best Actions},
    author={Weinzierl, Sven and Dunzer, Sebastian and Zilker, Sandra and Matzner, Martin},
    booktitle={Proceedings of the 18th International Conference on Business Process Management Forum},
    year={2020},
    publisher={Springer}
}

You can access the last paper here.

You can access the paper in which we presented the first prototype here.

We conducted all experiments on a workstation with 12 CPUs, 128 GB RAM. In Table 1, we present the times for training and testing of the baseline and our prescriptive business process monitoring technique.

Table 1: Times for training and testing (in seconds).

We implemented our technique in Python 3.6.8 and used the deep-learning framework TensorFlow 2.3 to build deep learning models.