---
title: "Completion Time and Next Activity Prediction of Processes Using Sequential Pattern Mining"
type: source
tags: [ppm, sequential-pattern-mining, next-activity, completion-time]
authors: [Ceci Michelangelo; Lanotte Pasqua Fabiana; Fumarola Fabio; Cavallo Dario Pietro; Malerba Donato]
year: 2014
venue: "Discovery Science (DS) 2014, LNAI 8777, Springer: 49–61"
kind: paper
raw_path: "raw/Predictive process monitoring/Completion Time and Next Activity Prediction of Processes Using Sequential Pattern Mining.pdf"
created: 2026-04-13
updated: 2026-04-13
key_claims:
  - Partial process models can be identified via sequential-pattern mining and used to train nested prediction models.
  - Event attribute information improves the model without degrading runtime.
  - Outperforms ProM5's transition-system approach for completion-time prediction; competitive for next-activity.
---

# Ceci et al. 2014 — Sequential Pattern Mining for PPM

Pre-deep-learning PPM that identifies **partial process models** via sequential pattern mining (an established data-mining technique), then trains nested prediction models on top. Dual target: completion time + next activity.

## Position in the landscape
Represents the **classical symbolic era** of PPM (Markov chains, transition systems, pattern mining, decision trees) that the deep-learning papers of 2017–2020 largely displaced for accuracy reasons, but which remain relevant for interpretability and small-data regimes.

## Connections
**Concepts:** [[concepts/predictive-process-monitoring]] · [[concepts/next-activity-prediction]] · [[concepts/remaining-time-prediction]]