Prabhakar Dixit

Detection and Interactive Repair of Event Ordering Imperfection in Process Logs

Many forms of data analysis require timestamp information to order the occurrences of events. The process mining discipline uses historical records of process executions, called event logs, to derive insights into business process behaviours and performance. Events in event logs must be ordered, typically achieved using timestamps. The importance of timestamp information means that it needs to be of high quality. To the best of our knowledge, no(semi-)automated support exists for detecting and repairing ordering-related imperfection issues in event logs. We describe a set of timestamp-based indicators for detecting event ordering imperfection issues in a log and our approach to repairing identified issues using domain knowledge. Lastly, we evaluate our approach implemented in the open-source process mining framework, ProM, using two publicly available logs.

Enabling interactive process analysis with process mining and visual analytics

In a typical healthcare setting, specific clinical care pathways can be defined by the hospitals. Process mining provides a way of analyzing the care pathways by analyzing the event data extracted from the hospital information systems. Process mining can be used to optimize the overall care pathway, and gain interesting insights into the actual execution of the process, as well as to compare the expectations versus the reality. In this paper, a generic novel tool called InterPretA, is introduced which builds upon pre-existing process mining and visual analytics techniques to enable the user to perform such process oriented analysis. InterPretA contains a set of options to provide high level conformance analysis of a process from different perspectives. Furthermore, InterPretA enables detailed investigative analysis by letting the user interactively analyze, visualize and explore the execution of the processes from the data perspective.

Detecting Changes in Process Behavior Using Comparative Case Clustering

Real-life business processes are complex and often exhibit a high degree of variability. Additionally, due to changing conditions and circumstances, these processes continuously evolve over time. For example, in the healthcare domain, advances in medicine trigger changes in diagnoses and treatment processes. Case data (e.g. treating physician, patient age) also influence how processes are executed. Existing process mining techniques assume processes to be static and therefore are less suited for the analysis of contemporary, flexible business processes. This paper presents a novel comparative case clustering approach that is able to expose changes in behavior. Valuable insights can be gained and process improvements can be made by finding those points in time where behavior changed and the reasons why. Evaluation using both synthetic and real-life event data shows our technique can provide these insights.

Using Domain Knowledge to Enhance Process Mining Results

Process discovery algorithms typically aim at discovering process models from event logs. Most algorithms achieve this by solely using an event log, without allowing the domain expert to influence the discovery in any way. However, the user may have certain domain expertise which should be exploited to create better process models. In this paper, we address this issue of incorporating domain knowledge to improve the discovered process model. First, we present a verification algorithm to verify the presence of certain constraints in a process model. Then, we present three modification algorithms to modify the process model. The outcome of our approach is a Pareto front of process models based on the constraints specified by the domain expert and common quality dimensions of process mining.

Fast Incremental Conformance Analysis for Interactive Process Discovery

Interactive process discovery allows users to specify domain knowledge while discovering process models with the help of event logs. Typically the coherence of an event log and a process model is calculated using conformance analysis. Many state-of-the-art conformance techniques emphasize on the correctness of the results, and hence can be slow, impractical and undesirable in interactive process discovery setting, especially when the process models are complex. In this paper, we present a framework (and its application) to calculate conformance fast enough to guide the user in interactive process discovery. The proposed framework exploits the underlying techniques used for interactive process discovery in order to incrementally update the conformance results. We trade the accuracy of conformance for performance. However, the user is also provided with some diagnostic information, which can be useful for decision making in an interactive process discovery setting. The results show that our approach can be considerably faster than the traditional approaches and hence better suited in an interactive setting.

Incremental Computation of Synthesis Rules for Free-Choice Petri Nets

In this paper, we propose a novel approach that calculates all the possible applications of synthesis rules, for well-formed free-choice Petri nets [8], in a speedy way to enable an interactive editing system. The proposed approach uses a so-called incremental synthesis structure, which can be used to extract all the synthesis rules, corresponding to a given net. Furthermore, this structure is updated incrementally, i.e. after usage of a synthesis rule, to obtain the incremental synthesis structure of the newly synthesized net. We prove that the proposed approach is correct and complete in order to synthesize any well-formed free-choice Petri net, starting with an initial well-formed atomic net and the corresponding incremental synthesis structure. A variant of the proposed approach has been implemented that allows interactive modeling (discovery) of sound business processes (from event logs). Experimental results show that the proposed approach is fast, and outperforms the baseline, and hence is well-suited for enabling interactive synthesis of very large nets.

Interactive Data-Driven Process Model Construction

Process discovery algorithms address the problem of learning process models from event logs. Typically, in such settings a user’s activity is limited to configuring the parameters of the discovery algorithm, and hence the user expertise/domain knowledge can not be incorporated during traditional process discovery. In a setting where the event logs are noisy, incomplete and/or contain uninteresting activities, the process models discovered by discovery algorithms are often inaccurate and/or incomprehensible. Furthermore, many of these automated techniques can produce unsound models and/or cannot discover duplicate activities, silent activities etc. To overcome such shortcomings, we introduce a new concept to interactively discover a process model, by combining a user’s domain knowledge with the information from the event log. The discovered models are always sound and can have duplicate activities, silent activities etc. An objective evaluation and a case study shows that the proposed approach can outperform traditional discovery techniques.