Petri Net Classes for Collaboration Mining: Assessment and Design Guidelines

TL;DR

This paper evaluates Petri net classes as potential standard models for representing diverse collaboration processes in mining, aiming to unify and improve collaboration discovery techniques.

Contribution

It introduces a meta model to assess Petri net classes for collaboration processes and provides design guidelines for collaboration discovery.

Findings

Petri net classes can model intra-organizational and cross-organizational collaborations.

A meta model helps evaluate Petri net suitability for diverse collaboration types.

Design guidelines support the development of effective collaboration mining techniques.

Abstract

Collaboration mining develops discovery, conformance checking, and enhancement techniques for collaboration processes. The collaboration process model is key to represent the discovery result. As for process mining in general, Petri Net classes are candidates for collaboration process models due to their analytical power. However, a standard model class to represent collaboration processes is lacking due to the heterogeneity of collaboration and, thus, of collaboration mining techniques. Collaboration heterogeneity requires to cover, for example, intra-organizational collaborations as well as choreographies that span a process across multiple organizations. A standard collaboration model class would advance collaboration mining by focusing discovery through a common target model, supporting comparison, and enabling flexible mining pipelines. To find a standard model class, we aim at capturing collaboration heterogeneity in a meta model, assess Petri net classes as candidates for collaboration process models through the meta model, and derive design guidelines for the collaboration discovery.