F Felix Mannhardt

From Low-Level Events to Activities - A Pattern-Based Approach

Process mining techniques analyze processes based on event data. A crucial assumption for process analysis is that events correspond to occurrences of meaningful activities. Often, low-level events recorded by information systems do not directly correspond to these. Abstraction methods, which provide a mapping from the recorded events to activities recognizable by process workers, are needed. Existing supervised abstraction methods require a full model of the entire process as input and cannot handle noise. This paper proposes a supervised abstraction method based on behavioral activity patterns that capture domain knowledge on the relation between activities and events. Through an alignment between the activity patterns and the low-level event logs an abstracted event log is obtained. Events in the abstracted event log correspond to instantiations of recognizable activities. The method is evaluated with domain experts of a Norwegian hospital using an event log from their digital whiteboard system. The evaluation shows that state-of-the art process mining methods provide valuable insights on the usage of the system when using the abstracted event log, but fail when using the original lower level event log.

On the fragmentation of process information : challenges, solutions, and outlook

An organization’s knowledge on its business processes represents valuable corporate knowledge because it can be used to enhance the performance of these processes. In many organizations, documentation of process knowledge is scattered around various process information sources. Such information fragmentation poses considerable problems if, for example, stakeholders wish to develop a comprehensive understanding of their operations. The existence of efficient techniques to combine and integrate process information from different sources can therefore provide much value to an organization. In this work, we identify the general challenges that must be overcome to develop such techniques. This paper illustrates how these challenges should be and, to some extent, are being met in research. Based on these insights, we present three main frontiers that must be further expanded to successfully counter the fragmentation of process information in organizations.

Decision Mining Revisited - Discovering Overlapping Rules

Decision mining enriches process models with rules underlying decisions in processes using historical process execution data. Choices between multiple activities are specified through rules defined over process data. Existing decision mining methods focus on discovering mutually-exclusive rules, which only allow one out of multiple activities to be performed. These methods assume that decision making is fully deterministic, and all factors influencing decisions are recorded. In case the underlying decision rules are overlapping due to non-determinism or incomplete information, the rules returned by existing methods do not fit the recorded data well. This paper proposes a new technique to discover overlapping decision rules, which fit the recorded data better at the expense of precision, using decision tree learning techniques. An evaluation of the method on two real-life data sets confirms this trade off. Moreover, it shows that the method returns rules with better fitness and precision in under certain conditions.

Data-driven process discovery : revealing conditional infrequent behavior from event logs

Process discovery methods automatically infer process models from event logs. Often, event logs contain so-called noise, e.g., infrequent outliers or recording errors, which obscure the main behavior of the process. Existing methods filter this noise based on the frequency of event labels: infrequent paths and activities are excluded. However, infrequent behavior may reveal important insights into the process. Thus, not all infrequent behavior should be considered as noise. This paper proposes the Data-aware Heuristic Miner (DHM), a process discovery method that uses the data attributes to distinguish infrequent paths from random noise by using classification techniques. Data- and control-flow of the process are discovered together. We show that the DHM is, to some degree, robust against random noise and reveals data-driven decisions, which are filtered by other discovery methods. The DHM has been successfully tested on several real-life event logs, two of which we present in this paper.

Extending Process Logs with Events from Supplementary Sources

Since organizations typically use more than a single IT system, information about the execution of a process is rarely available in a single event log. More commonly, data is scattered across different locations and unlinked by common case identifiers. We present a method to extend an incomplete main event log with events from supplementary data sources, even though the latter lack references to the cases recorded in the main event log. We establish this correlation by using the control-flow, time, resource, and data perspectives of a process model, as well as alignment diagnostics. We evaluate our approach on a real-life event log and discuss the reliability of the correlation under different circumstances. Our evaluation shows that it is possible to correlate a large portion of the events by using our method.

Enhancing process models to improve business performance : a methodology and case studies

Process mining is not only about discovery and conformance checking of business processes. It is also focused on enhancing processes to improve the business performance. While from a business perspective this third main stream is definitely as important as the others if not even more, little research work has been conducted. The existing body of work on process enhancement mainly focuses on ensuring that the process model is adapted to incorporate behavior that is observed in reality. It is less focused on improving the performance of the process. This paper reports on a methodology that creates an enhanced model with an improved performance level. The enhancements of the model limit incorporated behavior to only those parts that do not violate any business rules. Finally, the enhanced model is kept as close to the original model as possible. The practical relevance and feasibility of the methodology is assessed through two case studies. The result shows that the process models improved through our methodology, in comparison with state-of the art techniques, have improved KPI levels while still adhering to the desired prescriptive model.

Measuring the precision of multi-perspective process models

Process models need to reflect the real behavior of an organization’s processes to be beneficial for several use cases, such as process analysis, process documentation and process improvement. One quality criterion for a process model is that they should precise and not express more behavior than what is observed in logging data. Existing precision measures for process models purely focus on the control-flow dimension of a process model, thereby ignoring other perspectives, such as the data objects manipulated by the process, the resources executing process activities, and time-related aspects (e.g., activity deadlines). Focusing on the control-flow only, the results may be misleading. This paper extends existing precision measures to incorporate the other perspectives and, through an evaluation with a real-life process and corresponding logging data, demonstrates how the new measure matches our intuitive understanding of precision.

Revealing Work Practices in Hospitals Using Process Mining.

In order to improve health care processes (both in terms of quality and efficiency), we do need insight into how these processes are actually executed in reality. Interviewing health personnel and observing them in their work, are proven field-work techniques for gaining this insight. In this paper, we will introduce a complementary technique. This technique, called process mining, is based on the automatic analysis of digital events, registered in different information systems that support clinical work. Based on an event log, process mining can help in constructing a model of the process (discovery) or with checking to which extend an actual process confirms to a prescriptive model of it (conformance). This paper will briefly discuss two examples, which illustrate the use of process mining.

Guided process discovery : a pattern-based approach

Abstract Process mining techniques analyze processes based on events stored in event logs. Yet, low-level events recorded by information systems may not directly match high-level activities that make sense to process stakeholders. This results in discovered process models that cannot be easily understood. To prevent such situations from happening, low-level events need to be translated into high-level activities that are recognizable by stakeholders. This paper proposes the Guided Process Discovery method (GPD). Low-level events are grouped based on behavioral activity patterns, which capture domain knowledge on the relation between high-level activities and low-level events. Events in the resulting abstracted event log correspond to instantiations of high-level activities. We validate process models discovered on the abstracted event log by checking conformance between the low-level event log and an expanded model in which the high-level activities are replaced by activity patterns. The method was tested using two real-life event logs. We show that the process models discovered with the GPD method are more comprehensible and can be used to answer process questions, whereas process models discovered using standard process discovery techniques do not provide the insights needed.