Download This Paper Open PDF in Browser Add Paper to My Library Share: Permalink Using these links will ensure access to this page indefinitely Copy URL Copy DOI

In recent years, the industry has been witnessing an extended usage of process mining and automated event data analysis. Consequently, there is a rising significance in addressing privacy apprehensions related to the inclusion of sensitive and private information within event data utilized by process mining algorithms. State-of-the-art research mainly focuses on providing quantifiable privacy guarantees, e.g., via differential privacy, for trace variants that are used by the main process mining techniques, e.g., process discovery. However, privacy preservation techniques designed for the release of trace variants are still insufficient to meet all the demands of industry-scale utilization. Moreover, ensuring privacy guarantees in situations characterized by a high occurrence of infrequent trace variants remains a challenging endeavor. In this paper, we introduce two novel approaches for releasing differentially private trace variants based on trained generative models. With TraVaG, we leverage Generative Adversarial Networks (GANs) to sample from a privatized implicit variant distribution. Our second method employs Denoising Diffusion Probabilistic Models that reconstruct artificial trace variants from noise via trained Markov chains. Both methods offer industry-scale benefits and elevate the degree of privacy assurances, particularly in scenarios featuring a substantial prevalence of infrequent variants. Also, they overcome the shortcomings of conventional privacy preservation techniques, such as bounding the length of variants and introducing fake variants. Experimental results on real-life event data demonstrate that our approaches surpass state-of-the-art techniques in terms of privacy guarantees and utility preservation.

In higher education, data is collected that indicate the term(s) that a course is taken and when it is passed. Often, study plans propose a suggested course order to students. Study planners can adjust these based on detected deviations between the proposed and actual order of the courses being taken. In this work, we detect deviations by combining (1) the deviation between the proposed and actual course order with (2) the temporal difference between the expected and actual course-taking term(s). Partially ordered alignments identify the deviations between the proposed and actual order. We compute a partial order alignment by modeling a study plan as a process model and a student's course-taking behavior as a partial order. Using partial orders in such use cases allows one to relax the constraints of strictly ordered traces. This makes our approach less prone to the order in which courses are offered. Further, when modeling course-taking behavior as partial orders, we propose distinguishing intended course-taking behavior from actual course-passing behavior of students by including either all terms in which a course is attempted or only the term that a course is passed, respectively. This provides more perspectives when comparing the proposed and actual course-taking behavior. The proposed deviation measuring approach is evaluated on real-life data from RWTH Aachen University.

In business processes, the behavior, evolution and interactions of objects influence the outcome of process instances, and thus the value that a business user may assign to them. For example, in an order-to-cash process, a complete and timely delivery of a package is desirable, but depends on what happens to other objects upstream, like production batches. Negative outcomes call for a Root Cause Analysis (RCA) on the process. While many approaches for RCA using process mining exist, none is native to object-centric frameworks and thus suitable for capturing dependencies across object types. This work presents a method for RCA that operates on object-centric event logs (OCELs). Given an OCEL, our method returns a set of association rules on the activity level. These rules associate descriptive patterns over the various object types occurring at events with patterns indicating the process outcome. The patterns are abstracted from the log with the help of a first-order logic based query engine. A case study confirmed that our method can identify problematic interactions across various object types in real-life business processes.

Conformance checking is a crucial aspect of process mining, where the main objective is to compare the actual execution of a process, as recorded in an event log, with a reference process model, e.g., in the form of a Petri net or a BPMN. Conformance checking enables identifying deviations, anomalies, or non-compliance instances. It offers different perspectives on problems in processes, bottlenecks, or process instances that are not compliant with the model. Performing conformance checking in federated (inter-organizational) settings allows organizations to gain insights into the overall process execution and to identify compliance issues across organizational boundaries, which facilitates process improvement efforts among collaborating entities. In this paper, we propose a privacy-aware federated conformance-checking approach that allows for evaluating the correctness of overall cross-organizational process models, identifying miscommunications, and quantifying their costs. For evaluation, we design and simulate a supply chain process with three organizations engaged in purchase-to-pay, order-to-cash, and shipment processes. We generate synthetic event logs for each organization as well as the complete process, and we apply our approach to identify and evaluate the cost of pre-injected miscommunications.

In this study, we examine how event data from campus management systems can be used to analyze the study paths of higher education students. The main goal is to offer valuable guidance for their study planning. We employ process and data mining techniques to explore the impact of sequences of taken courses on academic success. Through the use of decision tree models, we generate data-driven recommendations in the form of rules for study planning and compare them to the recommended study plan. The evaluation focuses on RWTH Aachen University computer science bachelor program students and demonstrates that the proposed course sequence features effectively explain academic performance measures. Furthermore, the findings suggest avenues for developing more adaptable study plans.

We show that conducting a process-mining-centric analysis concerning cyber-physical systems provides insights into usage behavior. To show that, we perform our analysis on connected-vehicle data. We transform connected-vehicle data into an event log. We analyze the resulting event log using various process-mining techniques. In particular, we apply basic statistical analysis as well as process-discovery and conformance-checking techniques to receive a well-representative process model. We apply various process-enhancement techniques to get deeper insights. Finally, we capture a multi-perspective view using a state-based approach. We show deviations between a de-jure model and our picked process model, leading to better knowledge concerning real user behavior. We observed that the predefined escalation of warning states does not happen. Additionally, we verified system requirements. Furthermore, we show that the reasons for drivers’ behavior are not related to system issues. Applying process-mining techniques to data concerning cyber-physical systems provides valuable insights into their functionality in a real-world setting. By utilizing process-mining techniques, we can extract insights to a human-understandable level and provide a well-studied access point.

Process mining is rapidly growing in the industry. Consequently, privacy concerns regarding sensitive and private information included in event data, used by process mining algorithms, are becoming increasingly relevant. State-of-the-art research mainly focuses on providing privacy guarantees, e.g., differential privacy, for trace variants that are used by the main process mining techniques, e.g., process discovery. However, privacy preservation techniques for releasing trace variants still do not fulfill all the requirements of industry-scale usage. Moreover, providing privacy guarantees when there exists a high rate of infrequent trace variants is still a challenge. In this paper, we introduce TraVaG as a new approach for releasing differentially private trace variants based on Generative Adversarial Networks (GANs) that provides industry-scale benefits and enhances the level of privacy guarantees when there exists a high ratio of infrequent variants. Moreover, TraVaG overcomes shortcomings of conventional privacy preservation techniques such as bounding the length of variants and introducing fake variants. Experimental results on real-life event data show that our approach outperforms state-of-the-art techniques in terms of privacy guarantees, plain data utility preservation, and result utility preservation.

Simulation is a powerful tool to explore and analyze business processes and their potential improvements. Recorded event data allow for the generation of data-driven simulation models using process mining. The accuracy of existing approaches, however, remains a challenge. Various efforts are being made to improve the quality of the used data and techniques, such as extracting detailed resource performance. One of the least addressed challenges is the initial state of the simulation run. Starting from a steady state has been considered in simulation in other fields. In current process simulation approaches, the executions mostly start from an empty state. This assumption leads to initialization bias, or the startup problem, which has an impact on the early results and limits the types of analysis that can be performed. In this paper, we propose an approach to estimate a steady state of simulation models, which enables the generation of more realistic simulation results. The evaluation using real-world and synthetic event data shows the requirements for and advantages of starting from representative steady states in process simulations.

Recorded event data of processes inside organizations is a valuable source for providing insights and information using process mining. Most techniques analyze process executions at detailed levels, e.g., process instances, which may result in missing insights. Techniques at detailed levels using detailed event data should be complemented by techniques at aggregated levels. We designed and developed a standalone tool for diagnostics in event data of business processes based on both detailed and aggregated data and techniques. The data-driven framework first analyzes the event data of processes for possible compliance and performance problems, e.g., bottlenecks in processes. The results are used for aggregating the event data per window of time, i.e., extracting features in the time series format. The tool is able to uncover hidden insights in an explainable manner using time series analysis. The focus of the tool is to provide a data-driven business process analysis at different levels while reducing the dependencies on the user's domain knowledge for interpretation and feature engineering steps. The tool is applied to both real-world and synthetic event data.

Process simulation assesses the impact of changing environmental parameters on a process. To obtain realistic simulation models, process mining techniques can be deployed for a log-based discovery. Such discovery techniques usually rely on a fixed case notion, falling short in capturing the entangled nature of real organizational processes as an interplay of objects and subprocesses. Yet there is a need for such methods, given the requirement for information systems to foresee and adapt to changing environments in an online setting and in a holistic manner. In this work, we approach this research need by elaborating a method for simulation model discovery that is based on the object-centricity paradigm. To implement object-centric simulation, some intrinsic challenges have to be overcome. These include, first, the parametrizable generation of sets of objects having predefined interrelations that structure possible behavior. Second, the generated objects have to be synchronized and routed through a control-flow model. We outline these challenges, describe our solution approach, and evaluate the quality of both object generation and behavior.

AbstractIn short-term production management of the Internet of Production (IoP) the vision of a Production Control Center is pursued, in which interlinked decision-support applications contribute to increasing decision-making quality and speed. The applications developed focus in particular on use cases near the shop floor with an emphasis on the key topics of production planning and control, production system configuration, and quality control loops.Within the Predictive Quality application, predictive models are used to derive insights from production data and subsequently improve the process- and product-related quality as well as enable automated Root Cause Analysis. The Parameter Prediction application uses invertible neural networks to predict process parameters that can be used to produce components with desired quality properties. The application Production Scheduling investigates the feasibility of applying reinforcement learning to common scheduling tasks in production and compares the performance of trained reinforcement learning agents to traditional methods. In the two applications Deviation Detection and Process Analyzer, the potentials of process mining in the context of production management are investigated. While the Deviation Detection application is designed toidentify and mitigate performance and compliance deviations in production systems, the Process Analyzer concept enables the semi-automated detection of weaknesses in business and production processes utilizing event logs.With regard to the overall vision of the IoP, the developed applications contribute significantly to the intended interdisciplinary of production and information technology. For example, application-specific digital shadows are drafted based on the ongoing research work, and the applications are prototypically embedded in the IoP.

Different approaches are proposed for simulating processes in process mining. There are open challenges while designing the simulation models of processes: (1) the quality of the designed models is mostly evaluated using simulation results, and the models themselves do not get validated, (2) the choice of process aspects to be considered in the simulation models of processes as simulation parameters is rather arbitrary, e.g., considering multitasking, and (3) the distinction between the acquiring simulation parameters step and the parameters' regeneration step is not defined. This paper aims to introduce a reference meta-model for simulation in process mining. We derive the meta-model using the provided insights from process mining and the required parameters from the simulation techniques for simulating processes, i.e., Discrete Event Simulation (DES). This model enables the creation of process simulations and the comparison of approaches in relation to the process aspects under consideration. We illustrate the use of the model in practice by developing an automatic simulation model generation approach based on the reference model.

Process mining techniques are able to describe and model real processes using historic event data extracted from the information systems of organizations.Later, these insights are used for process improvement.For instance, Discrete Event Simulation (DES) uses process models that are able to mimic real-world events.However, the aggregated performance status of processes over time reveals various hidden relationships between process variables.Coarse-grained process logs are sets of performance variables over steps of time, generated using event data from processes.The coarse-grained process logs describe processes at higher levels.System Dynamics completes process mining by capturing the relationships between various process variables at a higher level of abstraction.In this paper, we propose a new framework for capturing conceptual models of processes using transformed event data.The main idea is to automatically discover the underlying relations as equations.This allows us to generate system dynamics simulations of processes.We employ a variety of statistical and machine learning techniques to discover the hidden relationships between process variables.The framework supports the simulation modeling task in the context of system dynamics simulations.The experiments using real event logs demonstrate that our approach is able to generate valid models and capture the underlying relationships.

Event data provide the main source of information for analyzing and improving processes in organizations. Process mining techniques capture the state of running processes w.r.t. various aspects, such as activity-flow and performance metrics. The next step for process owners is to take the provided insights and turn them into actions in order to improve their processes. These actions may be taken in different aspects of a process. However, simply being aware of the process aspects that need to be improved as well as potential actions is insufficient. The key step in between is to assess the outcomes of the decisions and improvements. In this paper, we propose a framework to systematically compare event data and the simulated event data of organizations, as well as comparing the results of modified processes in different settings. The proposed framework could be provided as an analytic service to enable organizations in easily accessing event data analytics. The framework is supported with a simulation tool that enables applying changes to the processes and re-running the process in various scenarios. The simulation step includes different perspectives of a process that can be captured automatically and modified by the user. Then, we apply a state-of-the-art comparison approach for processes using their event data which visually reflects the effects of these changes in the process, i.e., evaluating the process improvement. Our framework also includes the implementation of the change measurement module as a tool.

Process mining techniques make the underlying processes in organizations transparent. Historical event data are used to perform conformance checking and performance analyses. Analyzing a single process and providing visual insights has been the focus of most process mining techniques. However, comparing two processes or a single process in different situations is essential for process improvement. Different approaches have been proposed for process comparison. However, most of the techniques are either relying on the aggregated KPIs or their comparisons are based on process models, i.e., the flow of activities. Existing techniques are not able to provide understandable and insightful results for process owners. The current paper describes a tool that provides aggregated and detailed comparisons of two processes starting from their event logs using innovative visualizations. The visualizations provided by the tool are interactive. We exploit some techniques recently proposed in the literature, e.g., stochastic conformance checking and the performance spectrum, for conformance and performance comparison.

Abstract Den Kern des Exzellenzclusters Internet of Production bildet die domänen- und disziplinübergreifende Forschung in der Produktionstechnik. Der Fokus der Gruppe Short-Term Production Management liegt dabei insbesondere auf der Erhöhung von Entscheidungsqualität und -geschwindigkeit im Produktionsumfeld durch die datenbasierte Unterstützung der Anwender:innen. Dazu werden geeignete, kontextspezifische Daten aus Entwicklung, Produktion und Anwendung in Echtzeit und mit angemessener Granularität bereitgestellt, zusammengeführt und analysiert.

Process mining techniques transfer historical data of organizations into knowledge for the purpose of process improvement. Most of the existing process mining techniques are “backward-looking” and provide insights w.r.t. historical event data. Foreseeing the future of processes and capturing the effects of changes without applying them to the real processes are of high importance. Current simulation techniques that benefit from process mining insights are either at detailed levels, e.g., Discrete Event Simulation (DES), or at aggregated levels, e.g., System Dynamics (SD). System dynamics represents processes at a higher degree of aggregation and accounts for the influence of external factors on the process. In this paper, we propose an approach for simulating business processes that combines both types of data-driven simulation techniques to generate holistic simulation models of processes. These techniques replicate processes at various levels and for different purposes, yet they both present the same process. SD models are used for strategical what-if analysis, whereas DES models are used for operational what-if analysis. It is critical to consider the effects of strategical decisions on detailed processes. We introduce a framework integrating these two simulation models, as well as a proof of concept to demonstrate the approach in practice.

Process mining enables the discovery of actionable insights from event data of organizations. Process analysis techniques typically focus on process executions at detailed, i.e., fine-grained levels, which might lead to missed insights. For instance, the relation between the waiting time of process instances and the current states of the process including resources workload is hidden at fine-grained level analysis. We propose an approach for coarse-grained diagnostics of processes while decreasing user dependency and ad hoc decisions compared to the current approaches. Our approach begins with the analysis of processes at fine-grained levels focusing on performance and compliance and proceeds with an automated translation of processes to the time series format, i.e., coarse-grained process logs. We exploit time series analysis techniques to uncover the underlying patterns and potential causes and effects in processes. The evaluation using real and synthetic event logs indicates the efficiency of our approach to discover overlooked insights at fine-grained levels.

Process mining techniques use event data to describe business processes, where the provided insights are used for predicting processes’ future states ( Predictive Process Monitoring ). Remaining Time Prediction of process instances is an important task in the field of Predictive Process Monitoring (PPM). Existing approaches have two key limitations in developing Remaining Time Prediction Models (RTM): (1) The features used for predictions lack process context, and the created models are black-boxes. (2) The process instances are considered to be in isolation, despite the fact that process states, e.g., the number of running instances, influence the remaining time of a single process instance. Recent approaches improve the quality of RTMs by utilizing process context related to batching-at-end inter-case dynamics in the process, e.g., using the time to batching as a feature. We propose an approach that decreases the previous approaches’ reliance on user knowledge for discovering fine-grained process behavior. Furthermore, we enrich our RTMs with the extracted features for multiple performance patterns (caused by inter-case dynamics), which increases the interpretability of models. We assess our proposed remaining time prediction method using two real-world event logs. Incorporating the created inter-case features into RTMs results in more accurate and interpretable predictions.