Discovery of Improvement Opportunities in Knock-out Checks of Business Processes
Name
Lino Moises Mediavilla Ponce
Abstract
Business processes such as loan applications or university admissions usually contain a set of activities known as knock-out checks, which classify cases into two groups: accepted and rejected. When a knock-out check rejects a case, all the work previously performed on it is considered a waste. This waste can be reduced by modifying how the checks are performed. Previous studies have provided heuristics and tools for identifying such improvement opportunities from process models at design-time, while others have proposed predictive, black-box models for reordering knock-out checks at run-time. However, they have not provided a method for obtaining knock-out check insights and improvement opportunities directly from existing processes’ data. Here, we show a data-driven, decision-rules-based approach for discovering improvement opportunities related to the knock-out checks of business processes. Experiments on synthetic and real-world event logs show that the approach successfully identifies improvement opportunities while attaining a performance comparable to black-box approaches. Moreover, by leveraging interpretable Machine Learning techniques, our approach provides further insights into the knock-out checks of business processes that black-box approaches do not.
Graduation Thesis language
English
Graduation Thesis type
Master - Software Engineering
Supervisor(s)
Fredrik Milani, Manuel Camargo, Katsiaryna Lashkevich
Defence year
2022