Conformance Checking with Regulations

业务流程通常必须遵循特定的规定法规，如医疗保健的临床实践指南、公共管理的法律和法规，或许多不同领域的新卫生规则。对于组织的成功和正式的审计，必须知道：我们是否遵循规定的规定？如果我们偏离了方向，为什么？我们应该改进员工的培训吗？流程挖掘的一个主要好处是，它提供了对业务流程和基于证据的流程分析技术的真实执行的洞察。一致性检查是过程挖掘的一项主要任务，它包括检查设计的过程模型和实际行为之间的关系，以及识别和分析它们之间的偏差的技术。因此，它允许组织回答上述问题。然而，要实现工艺法规的符合性检查，需要解决主要挑战。首先，法规通常以冗长且模棱两可的文本形式编写，因此不能直接用于一致性检查。参考模型，即用于设计其他流程的流程模型模板，可以弥合法规和流程执行数据之间的这一差距。然而，将法规转换为参考模型需要大量的人工工作，而关于自动化这一步骤的研究仍处于起步阶段。其次，现代一致性检查技术在计算上是高效的，但无法判断偏差的相关性，无法考虑多个模型，也无法破译期望的(正的)和不期望的(负的)偏差，所有这些在处理法规时都很重要。此外，关于如何可视化一致性检查结果的研究非常有限，尤其是在目标用户是企业用户的情况下。最后，流程执行数据比参考模型详细得多，因此需要高度的自动化抽象来连接它们。Cher项目首次结合了参考建模和一致性检查的技术，以比较真实的流程行为和规定的规则。目标是发现并可视化它们之间的偏差，以便为员工提供量身定做的培训、准备审计或改进各自领域中的流程或法规的建议。需要针对几个开放的方面来允许与参考模型的一致性检查，包括(1)支持(半自动)参考模型的生成，(2)为这类流程挖掘项目提取有用的事件日志，(3)现有一致性检查方法的基准及其可能的扩展，以及(4)关于CHER方法如何允许利用符合法规的一致性检查的经验评估，例如用于培训员工。

Business processes often have to follow specific prescribed regulations, such as clinical practice guidelines in healthcare, laws and statutes in public administration, or the new hygiene rules in many different domains. For both organizational success and official audits, it is essential to know: Are we following the prescribed regulations? If we deviate, why? Should we improve employees’ training? Could the rules be adapted to be better applicable in the real world?A major benefit of process mining is that it provides insights into the real execution of business processes and techniques for evidence-based process analysis. Conformance checking, a main task of process mining, comprises techniques for checking the relation between a designed process model and the real-life behavior, and identify as well as analyze deviation between them. Hence, it allows organizations to answer the above-raised questions. However, to realize conformance checking on process regulations, major challenges need to be addressed. First, regulations are typically written as long and ambiguous texts, so they cannot be used directly for conformance checking. Reference models, i.e., process model templates that serve to be re-used for the design of other processes, can bridge this gap between regulations and process execution data. However, transforming regulations into reference models requires high manual effort, and research on automating this step is still at the beginning. Second, modern conformance checking techniques are computationally efficient, but unable to judge the deviation’s relevance, consider multiple models, or decipher desirable (positive) from undesirable (negative) deviations, all of which are important when dealing with regulations. In addition, there is only limited research on how to visualize conformance checking results, particularly if the intended audience is business users. Finally, process execution data is much more detailed than reference models, such that a high degree of automated abstraction is required to connect them.The CheR project combines, for the first time, techniques from reference modeling and conformance checking to compare real-life process behavior with prescribed regulations. The goal is to find and visualize the deviations between them to allow tailored training for employees, preparation of audits, or suggestions for improving either the process or the regulations in the respective domains. Several open aspects need to be targeted to allow conformance checking with reference models, including (1) supporting (semi)-automatic generation of reference models, (2) the extraction of useful event logs for this type of process mining project, (3) benchmarking of existing conformance checking methods and their possible extension, and (4) an empirical evaluation on how the CheR approach allows to leverage conformance checking with regulations, e.g., for training employees.