---
title: "Event-based Failure Prediction in Distributed Business Processes"
type: source
tags: [ppm, failure-prediction, distributed, choreography, ml]
authors: [Borkowski Michael; Fdhila Walid; Nardelli Matteo; Rinderle-Ma Stefanie; Schulte Stefan]
year: 2018
venue: "Information Systems (Elsevier), DOI:10.1016/j.is.2017.12.005"
kind: paper
raw_path: "raw/Predictive process monitoring/Event based failure prediction in distributed business processes Borkowski 2017.pdf"
created: 2026-04-13
updated: 2026-04-13
key_claims:
  - Event-based failure prediction is needed for highly distributed inter-organisational choreographies, where rule-based prediction breaks down.
  - Machine learning offers a pragmatic alternative to rule-based failure prediction.
  - Evaluated on real-world business-process test data.
---

# Borkowski et al. 2018 — Event-based Failure Prediction in Distributed Processes

Targets **distributed / choreographic** business processes — cross-organisational settings where no central engine observes all events and rule-based failure prediction is impractical.

## Contribution
- Frames failure prediction over **event streams** from heterogeneous actors.
- Demonstrates ML as a scalable alternative to rule-based prediction.
- Evaluates on a real-world cross-organisation test dataset.

## Significance
Connects PPM to the choreography / inter-organisational dimension that the APM Manifesto identifies as a key research direction ("processes rarely exist in isolation" — see [[sources/2026-calvanese-agentic-bpm-manifesto]] challenge **M3**).

## Connections
**Concepts:** [[concepts/predictive-process-monitoring]] · [[concepts/outcome-prediction]]
**Authors:** [[entities/stefanie-rinderle-ma]]