No KnowledgeWithout Processes Process Mining as a Tool to Find Out What People and Organizations Really Do

摘要

In recent years, process mining emerged as a new and exciting collection of analysis approaches. Process mining combines process models and event data in various novel ways. As a result, one can find out what people and organizations really do. For example, process models can be automatically discovered from event data. Compliance can be checked by confronting models with event data. Bottlenecks can be uncovered by replaying timed events on discovered or normative models. Hence, process mining can be used to identify and understand bottlenecks, inefficiencies, deviations, and risks. Despite the many successful applications of process mining, few people are aware of the recent advances in process mining. One of the main reasons is that process mining is not part of existing (a) data mining, (b) machine learning, (c) business intelligence, (d) process modeling, and (e) simulation approaches and tools. For example, conventional "data miners" use a very broad definition of data mining, but at the same time focus on a limited set of classical problems unrelated to process models (e.g., decision tree learning, regression, pattern mining, and clustering). None of the classical data mining tools supports process mining techniques such as process discovery, conformance checking, and bottleneck analysis. This keynote paper briefly summarizes the differences between process mining and more established analysis and modeling approaches. Moreover, the paper emphasis the need to extract process-related knowledge.