ModErARe - Modeling Error Analysis and Resolution

Although process modeling has gained increasing importance for documenting business operations and automating workflow execution, process models still display a wide range of quality problems impeding their comprehensibility and consequently hampering their maintainability. Literature reports, for example, on error rates between 10% and 20% in industrial process model collections. These problems have resulted in vivid research on the quality of process models with the goal of obtaining a better understanding of factors influencing the quality of process models. Thereby, existing research mostly focuses on the product or outcome of process modeling. Recently, a new stream of research emerged that aims at obtaining a general understanding of the process followed to create process modelsthe process of process modeling (PPM). Even though it is known that quality issues frequently arise during the PPM, it is not clear at what point quality issues are introduced, how they can be discovered, and in what way they can be resolved by process modelers. The ModErARe project aims to close this research gap by systematically investigating quality issues that occur during the process of process modeling. More specifically, ModErARe investigates why quality issues occur, how quality issues are discovered, and how they are resolved by looking at the PPM. ModErARe not only provides a better understanding of typical quality issues during the PPM, but also of their occurrence (e.g., problem patterns frequently resulting in quality issues or reasons for quality issues). As a further outcome, ModErARe provides methods and techniques for predicting quality issues and hence for preventing them. In addition, enabled by better understanding of the processes involved in the discovery and resolution of quality issues, ModErARe contributes methods and techniques that provide guidance to process modelers during the PPM for discovering and resolving quality issues. Ultimately, this leads to improved modeling outcomes through error prevention as well as support for error discovery and resolution.

Although process modeling has gained increasing importance for documenting business operations and automating workflow execution, process models still display a wide range of quality problems impeding their comprehensibility and consequently hampering their maintainability. Literature reports, for example, on error rates between 10% and 20% in industrial process model collections. These problems have resulted in vivid research on the quality of process models with the goal of obtaining a better understanding of factors influencing the quality of process models. Thereby, existing research mostly focuses on the product or outcome of process modeling. Recently, a new stream of research emerged that aims at obtaining a general understanding of the process followed to create process models -- the process of process modeling (PPM). Even though it is known that quality issues frequently arise during the PPM, it is not clear at what point quality issues are introduced, how they can be discovered, and in what way they can be resolved by process modelers. The ModErARe project aimed to close this research gap by systematically investigating the quality issues that occur during the process of process modeling. More specifically we developed techniques for automatically assessing the evolution of different quality aspects over time and designed corresponding visualizations. Additionally, ModErARe investigated co-occurrences of issues using statistical methods with the goal to predict quality issues and developed an approach for expertise prediction based on the evolution of quality issues. Moreover, ModErARe contributed towards a better understanding of how humans inspect process models, what strategies are taken, what challenges arise, and what cognitive processes are involved. The results of ModErARe also opened new future research opportunities including the development of neuro-adaptive systems.