INEXA: Interactive and Explainable Process Model Abstraction Through Object-Centric Process Mining

Core Concepts

INEXA enables interactive exploration of process models at different granularity levels while maintaining a link to the event log.

Abstract

Process events recorded by multiple information systems at different granularity levels.
Existing process model abstraction techniques aim to reduce model size but may disconnect from the event log.
INEXA proposes an interactive, explainable process model abstraction method.
INEXA aggregates large process models to a displayable size while maintaining the link to the event log.
The method allows for interactive exploration of granularity levels with automatic tracing of applied abstractions.

Stats

"The discovered process model of a real-world manufacturing process consists of 1,489 model elements and over 2,000 arcs."
"The event log fragment L is based on a bank’s account opening business process."
"The event log consists of 9,460 events recorded by the Cloud Process Execution Engine during the production of chess tower pieces."

Quotes

"To enable the exploration of discovered process models at different granularity levels, we propose INEXA, an interactive, explainable process model abstraction method that keeps the link to the event log."
"Overall, the lack of existing methods for interactive, explainable process model abstraction raises the following research question:"

Key Insights Distilled From

INEXA

by Janik-Vasily... at arxiv.org 03-28-2024

https://arxiv.org/pdf/2403.18659.pdf

Deeper Inquiries

How can interactive and explainable process model abstraction methods benefit various industries?

Interactive and explainable process model abstraction methods can benefit various industries by providing a more intuitive and user-friendly way to explore and analyze complex process models. By allowing users to interactively adjust the granularity levels of the process model, these methods enable a more tailored and focused analysis, leading to better insights and decision-making. In industries such as manufacturing, healthcare, finance, and logistics, where processes can be highly complex and interconnected, interactive and explainable process model abstraction methods can help in identifying bottlenecks, inefficiencies, and opportunities for optimization. Additionally, these methods can enhance collaboration among stakeholders by providing a clear and understandable representation of the processes, facilitating communication and alignment on process improvements.

What are the potential limitations of maintaining a link to the event log in process model abstraction?

Maintaining a link to the event log in process model abstraction can have some limitations. One potential limitation is the increased complexity and computational overhead involved in tracking and updating the event log with each abstraction applied to the process model. This can lead to performance issues, especially when dealing with large event logs and complex process models. Additionally, the link to the event log may introduce dependencies between the process model and the underlying data, making it challenging to modify or update the process model without affecting the event log. Furthermore, ensuring the consistency and accuracy of the event log throughout the abstraction process can be a challenge, as any discrepancies or errors in the event log can impact the validity of the process model abstraction.

How can the concept of object-centric process mining be further expanded and applied in real-world scenarios?

The concept of object-centric process mining can be further expanded and applied in real-world scenarios by incorporating more diverse and specialized object types that reflect the specific characteristics of different industries and processes. By defining object types that capture not only the workflow of business objects but also resources, devices, subprocesses, and interactions, object-centric process mining can provide a more comprehensive and detailed view of the processes. This expanded approach can enable more accurate process discovery, analysis, and optimization, leading to improved operational efficiency, compliance, and decision-making in various industries. Additionally, integrating advanced data analytics techniques, such as machine learning and predictive modeling, with object-centric process mining can enhance the predictive capabilities and insights derived from process data, enabling organizations to proactively identify patterns, trends, and anomalies in their processes.