nesy4ppm 0.1

Python library to perform learning and suffix prediction with background knowledge under concept drifts.

NeSy4PPM documentation

NeSy4PPM is the first Python package designed for both single-attribute (e.g., activity) and multi-attribute (e.g., activity and resource) suffix prediction in predictive process monitoring. It implements a Neuro-Symbolic (NeSy) system that integrates neural models with various types of symbolic background knowledge (BK), enabling accurate and compliant predictions even under concept drift.

NeSy4PPM offers the following key features:

1. Symbolic knowledge integration: supports declarative and procedural BK, including DECLARE, MP-DECLARE (multi-perspective DECLARE), ProbDECLARE (probabilistic DECLARE), and Petri nets.

2. Flexible learning: provides multiple prefix encoding methods and supports LSTM (Long Short-Term Memory) and Transformer architectures.

3. Drift-aware prediction: contextualizes neural predictions using BK in real-time, enhancing prediction accuracy and compliance in dynamic environments.

Installation

We recommend using a virtual environment to avoid conflicts between your local Python setup and required libraries. You can use Conda or venv:

1. Clone/download the NeSy4PPM project.

2. Activate your virtual environment.

3. Install the dependencies listed in requirements.txt using:: pip install -r requirements.txt

Alternatively, you can install NeSy4PPM directly from PyPi.

For ProbDECLARE BK conformance checking, install the Lydia backend using Docker:

1. Install Docker

2. Pull the Lydia image:

   docker pull whitemech/lydia:latest

3. Make the Docker image executable under the name lydia. On Linux and macOS machines, the following commands should work:

   echo '#!/usr/bin/env sh' > lydia
   echo 'docker run -v$(pwd):/home/default whitemech/lydia lydia "$@"' >> lydia
   sudo chmod u+x lydia
   sudo mv lydia /usr/local/bin/

More information can be found at Logaut repository.

Tutorials

Here, you will find tutorials explaining how to utilize the different functionalities of NeSy4PPM. All tutorials can be downloaded and run from our Github. The tutorials use the Helpdesk log as a running example.

Repository Structure

• NeSy4PPM/learning: contains the implementation of event log encoding and Neural Networks model training.

• NeSy4PPM/ProbDeclmonitor: contains the implementation of suffix prediction using a contextualized Neural predictions with BK.

• tests/: provides evaluation scripts for assessing the NeSy4PPM prediction performance.

• docs/tutorials/: contains step-by-step guides and examples to help users get started with NeSy4PPM.