Timeline-based process discovery

Process mining techniques aim to analyze and improve conformance and performance of processes using event data. Process discovery is the most prominent process-mining task: A process model is derived based on an event log. The process model should be able to capture causalities, choices, concurrency, and loops. Process discovery is very challenging because of trade-offs between fitness, simplicity, precision, and generalization. Note that event logs typically only hold example behavior and cannot be assumed to be complete (to avoid overfitting). Dozens of process discovery techniques have been proposed. These use a wide range of approaches, e.g., language- or state-based regions, genetic mining, heuristics, expectation maximization, iterative log-splitting, etc. When models or logs become too large for analysis, the event log may be automatically decomposed or traces may be clustered before discovery. Clustering and decomposition are done automatically, i.e., no additional information is used. This paper proposes a different approach where a localized event log is assumed. Events are localized by assigning a non-empty set of regions to each event. It is assumed that regions can only interact through shared events. Consider for example the mining of software systems. The events recorded typically explicitly refer to parts of the system (components, services, etc.). Currently, such information is ignored during discovery. However, references to system parts may be used to localize events. Also in other application domains, it is possible to localize events, e.g., communication events in an organization may refer to multiple departments (that may be seen as regions). This paper proposes a generic process discovery approach based on localized event logs. The approach has been implemented in ProM and experimental results show that location information indeed helps to improve the quality of the discovered models.KeywordsVisible TraceConformance CheckVertical DecompositionInductive MinerUnique TraceThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Process mining, especially process discovery, has been utilized to extract process models from event logs. One challenge faced by process discovery is to identify concurrency effectively. State-of-the-art approaches employ activity orders in traces to undertake process discovery and they require stringent completeness notions of event logs. Thus, they may fail to extract appropriate processes when event logs cannot meet the completeness criteria. To address this problem, we propose in this paper a novel technique which leverages activity dependences in traces. Based on the observation that activities with no dependencies can be executed in parallel, our technique is in a position to discover processes with concurrencies even if the logs fail to meet the completeness criteria. That is, our technique calls for a weaker notion of completeness. We evaluate our technique through experiments on both real-world and synthetic event logs, and the conformance checking results demonstrate the effectiveness of our technique and its relative advantages compared with state-of-the-art approaches.

A well-known information reuse approach is to exploit event logs for process discovery and optimization. However, process discovery is rarely used for product development projects. This is because information systems in product development, e. g., Product-Lifecycle-Management (PLM) systems, do not provide the event logs required by process discovery algorithms. Additionally, existing algorithms struggle with development projects, as these are unstructured and rich in variety. In this paper, we propose a novel approach to process discovery in order to make it applicable and tailored to product development projects. Instead of using flat event logs, we provide a graph-based data structure that is able to represent both activities and data of product development projects with the dataflow between activities. Based on this structure, we can leverage provenance available in PLM systems. Furthermore, we may use frequent subgraph mining to discover process patterns. Such patterns are well suited to describe different variants and common sub-processes of unstructured processes. Using a prototype, we evaluate this approach and successfully discover prevailing patterns. These patterns may be used by engineers to support their decision-making or help improve the execution of development projects.

Massive amounts of business process event logs are collected and stored by modern information systems. Numerous process discovery approaches have been proposed to extract descriptive process models from such event logs in the past decades. To improve process discovery efficiency, event log sampling techniques are proposed. A sample log is a delicately selected subset of the original log that requires less computational cost. However, existing sampling techniques have difficulties, e.g., low efficiency, in handling large-scale event logs. To tackle this challenge, we propose a novel ranking-based event log sampling approach, denoted as \( LogRank^+ \), to support efficient sampling. In addition, we introduce a framework to evaluate the effectiveness of different sampling techniques by quantifying the sampling efficiency and the quality of sample logs. The proposed sampling approach has been implemented in the open-source process mining toolkit ProM. Experimental evaluation with both synthetic and real-life event logs demonstrates that the proposed sampling approach provides an effective solution to improve event log sampling efficiency as well as ensuring high quality of the obtained sample logs from a process discovery perspective.

In the domain of process discovery, there are four quality dimensions for evaluating process models of which simplicity is one. Simplicity is often measured using the size of a process model, the structuredness and the entropy. It is closely related to the process model understandability. Researchers from the domain of business process management (BPM) proposed several metrics for measuring the process model understandability. A part of these understandability metrics focus on the control-flow perspective, which is important for evaluating models from process discovery algorithms. It is remarkable that there are more of these metrics defined in the BPM literature compared to the number of proposed simplicity metrics. To research whether the understandability metrics capture more understandability dimensions than the simplicity metrics, an exploratory factor analysis was conducted on 18 understandability metrics. A sample of 4450 BPMN models, both manually modelled and artificially generated, is used. Four dimensions are discovered: token behaviour complexity, node IO complexity, path complexity and degree of connectedness. The conclusion of this analysis is that process analysts should be aware that the measurement of simplicity does not capture all dimensions of the understandability of process models.

Many real world business processes are executed without explicit orchestration and hence do not generate structured execution logs. This is particularly true for the class of business processes which are executed in service delivery centers in emerging markets where rapid changes in processes and in the people executing the processes are common. In such environments, the process execution logs are usually a mix of human entered activity log of actions performed and the auto-generated logs by various tools used during the process execution. Process discovery from unstructured execution logs has been a relatively unexplored research area. In this paper, we propose an approach for process discovery from unstructured logs using gaussian mixture models and hidden markov models. We apply this approach to the logs generated by a real-world business process used in a service delivery center and demonstrate that the results obtained are comparable to an approach of manually labeling the logs followed by a best known process discovery algorithm in literature. The approach proposed is generic and applicable to a wide range of business process execution settings.

Churning of the customer base is always a top issue in Banking. It is directly related to recurrent revenue, and the ever increasing acquisition costs for new customers. In a first approach, this issue is related to both the quality of service (which is mainly in the front-office, say the contact center) and the speed of service, which is mainly in the back-office. Many studies published to date on this required manual data collection. This creates in general two concerns: worker behavior may change under observation, and manual data collection is expensive and often error prone. In this paper it is shown by means of a case study for a Multi-National Bank (with 5000 employees in the back office) how automated Business Process Discovery, which is an advanced type of process mining, makes it possible to handle the above concerns. The automated data collection and the analysis, in terms of Hidden Markov Models, are key elements. Several results regarding the quality and speed of service have been obtained. Most interesting was the discovery of deeper root causes for customer attrition. Once the deficiencies in the processes are identified, appropriate process improvements can be designed and simulated based on the models emerging from process discovery. In this case study, significant quality and speed improvements as well as customer churn reductions have been obtained.

Process mining has emerged as a new scientific research topic on the interface between process modeling and event data gathering. In the search for process models that best fit to reality, the process discovery approach of creating referential processes from observed behavior. However, despite these methods showing relevant results, when faced with noisy and divergent tendencies they end up producing limited results. This work proposes the application of process discovery technique, combined to cluster technique k-means, to generate new process models, considering its conformance checking measures. The proposed solution is applied to an ad hoc workflow. And as a result, the use of the clustering techniques coupled with process discovery showed significant gains in the generation of process models, unlike the standard approach.

One of the most actively researched areas in the field of process mining is process discovery, which aims to construct a schema that aligns with existing event trace sequences. Current standard industrial workflow schema induction methods impose certain limitations on the system being examined. To address the shortcomings, this article proposes a novel solution that employs graph neural networks and convolutional neural networks to perform schema discovery. In the first phase of schema generation, we perform equivalence prediction, implemented as an edge prediction task. From the obtained equivalence network, we identify the target schema nodes, which correspond to the maximal quasi-cliques of this network. The results of the performed efficiency tests demonstrate that the proposed method can manage such complex cases that are not covered by standard process discovery methods, and it provides more compact and more precise schema graphs.

Process discovery aims at building process models using information retrieved from logs. Process characteristics play a significant role in the selection of a suitable process modeling language for describing process discovery results. Business processes characterized by high variability, in which participants have a lot of autonomy and flexibility in executing the process, are difficult to be described with procedural process modeling languages,since they explicitly represent in a model every possible path. Declarative languages, like Declare, alleviate this issue by defining a set of constraints between activities that must not be violated during the process execution instead of describing what to do step by step. Recently, several process discovery techniques have been proposed for extracting a set of Declare constraints from a log. However, no one of these techniques allows the user to exploit the time perspective often available in a log to discover “time-aware” Declare constraints. Timed Declare has already previously been introduced to monitor metric temporal constraints at runtime. In this paper, we use this semantics for discovering a set of Timed Declare constraints from an event log. We have implemented the proposed approach as a plug-in of the process mining tool ProM. We have validated the approach by using our plug-in to mine two real-life event logs.

In this work, we explore an approach to process discovery that is based on combining several existing process discovery algorithms. We focus on algorithms that generate process models in the process tree notation, which are sound by design. The main components of our proposed process discovery approach are the Inductive Miner, the Evolutionary Tree Miner, the Local Process Model Miner and a new bottom-up recursive technique. We conjecture that the combination of these process discovery algorithms can mitigate some of the weaknesses of the individual algorithms. In cases where the Inductive Miner results in overgeneralizing process models, the Evolutionary Tree Miner can often mine much more precise models. At the other hand, while the Evolutionary Tree Miner is computationally expensive, running it only on parts of the log that the Inductive Miner is not able to represent with a precise model fragment can considerably limit the search space size of the Evolutionary Tree Miner. Local Process Models and bottom-up recursion aid the Evolutionary Tree Miner further by instantiating it with frequent process model fragments. We evaluate our approaches on a collection of real-life event logs and find that it does combine the advantages of the miners and in some cases surpasses other discovery techniques.

Process discovery aims to derive a process model that accurately represents the observed behavior in an event log. As a state-of-the-art process discovery technique, Inductive Miner (IM) generates sound process models (i.e., free of deadlocks) while ensuring optimal replay fitness. However, IM may sometimes produce over-generalized process models with locally imprecise structures, often resulting in the creation of so-called flower structures. To address this limitation, this paper presents a novel technique that refines the process model generated by IM by optimizing its imprecise sub-processes. Specifically, the technique begins by identifying and extracting sub-logs corresponding to imprecise sub-processes in the initial IM-generated process model. Then, these imprecise sub-processes are iteratively optimized using a frequency-based filtering mechanism applied to the sub-logs. Once optimized, the imprecise sub-processes in the initial process model are replaced by the optimized ones, generating a set of candidates process models. Finally, the candidate with the best quality, in terms of fitness and precision, is selected as the final optimized process models. The proposed technique has been implemented as a plugin for the open-source process mining platform ProM. Through comparisons with state-of-the-art process discovery techniques using 10 publicly available real-life event logs, the experimental results demonstrate that the proposed method achieves an average absolute improvement of 0.173 in F-measure over its IMi variant, while also exhibiting competitive performance relative to other state-of-the-art approaches.

Microservice architecture is widely adopted in cloud computing for building large-scale distributed systems. However, due to frequent calls between microservices, such systems become increasingly complex and fragile, leading to reduced reliability of the system behavior. Therefore, implementing process discovery and constructing process models for the historical logs of microservice systems becomes crucial to analyzing system behavior. However, existing process discovery methods do not effectively capture the synchronous or asynchronous call relations between microservices at the endpoint level, limiting the ability to analyze the impact of microservice interactions on system performance. This paper proposes a novel process discovery capable of capturing call relations in microservice systems at the endpoint level. First, the call events related to the microservice calls are extracted from the interleaved fragmented raw logs by event serialization to form event sequences. Then, an invocation tree is defined and constructed, and pruning is applied to obtain the subsequence that can be used to derive call relations correctly. Based on the defined event relations, the call relations between endpoints are discovered. Finally, the <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$\alpha$</tex-math></inline-formula>-based miner is applied to perform the mining and generation of the process model to obtain the defined microservice call net. Experimental results show that, compared to state-of-the-art methods, the proposed approach can accurately and effectively build the microservice system process model.

Process mining practices are mainly activity-oriented and they seldom consider the (often conflicting) goals of stakeholders. Involving goal-related factors, as often done in requirements engineering, can improve the rationality and interpretability of mined models and lead to better opportunities to satisfy stakeholders. This paper proposes a new Goal-oriented Process Enhancement and Discovery (GoPED) method to align discovered models with stakeholders’ goals. GoPED first adds goal-related attributes to traditional event characteristics (case identifier, activities, and timestamps), selects a subset of cases with respect to a goal-related criterion, and finally discovers a process model from that subset. We define three types of criteria that suggest desired satisfaction levels from a (i) case perspective, (ii) goal perspective, and (iii) organization perspective. For each criterion, an algorithm is proposed to enable selecting the best subset of cases were the criterion holds. The resulting process models are expected to reproduce the desired level of satisfaction. A synthetic event log is used to illustrate the proposed algorithms and to discuss their results.

Compliance in business processes has become a fundamental requirement given the constant rise in regulatory requirements and competitive pressures that have emerged in recent decades. While in other areas of business process modelling and execution, considerable progress towards automation has been made (e.g., process discovery, executable process models), the interpretation and implementation of compliance requirements is still a highly complex task requiring human effort and time. To increase the level of “mechanization” when implementing regulations in business processes, compliance research seeks to formalize compliance requirements. Formal representations of compliance requirements should, then, be leveraged to design correct process models and, ideally, would also serve for the automated detection of violations. To formally specify compliance requirements, however, multiple process perspectives, such as control flow, data, time and resources, have to be considered. This leads to the challenge of representing such complex constraints which affect different process perspectives. To this end, current approaches in business process compliance make use of a varied set of languages. However, every approach has been devised based on different assumptions and motivating scenarios. In addition, these languages and their presentation usually abstract from real-world requirements which often would imply introducing a substantial amount of domain knowledge and interpretation, thus hampering the evaluation of their expressiveness. This is a serious problem, since comparisons of different formal languages based on real-world compliance requirements are lacking, meaning that users of such languages are not able to make informed decisions about which language to choose. To close this gap and to establish a uniform evaluation basis, we introduce a running example for evaluating the expressiveness and complexity of compliance rule languages. For language selection, we conducted a literature review. Next, we briefly introduce and demonstrate the languages’ grammars and vocabularies based on the representation of a number of legal requirements. In doing so, we pay attention to semantic subtleties which we evaluate by adopting a normative classification framework which differentiates between different deontic assignments. Finally, on top of that, we apply Halstead’s well-known metrics for calculating the relevant characteristics of the different languages in our comparison, such as the volume, difficulty and effort for each language. With this, we are finally able to better understand the lexical complexity of the languages in relation to their expressiveness. In sum, we provide a systematic comparison of different compliance rule languages based on real-world compliance requirements which may inform future users and developers of these languages. Finally, we advocate for a more user-aware development of compliance languages which should consider a trade off between expressiveness, complexity and usability.