1.    Introduction

In the current business landscape, organizations extensively depend on various information systems to support their operations, especially their business processes (Revina & Aksu, 2023). Business processes are crucial as they have a direct or indirect impact on the delivery of services to customers, thereby significantly influencing levels of customer satisfaction (Martin et al., 2021).

A pioneering approach in the field of business process management is Process Mining (PM). Unlike data mining, which focuses on analyzing data, process mining adopts a process-centric approach. It utilizes specialized software known as business mining applications to discover, evaluate, or enhance business processes within organizations. Traditionally, process mining methods have utilized digital logs as the primary source for analysis (van der Aalst, 2012). However, over the past two decades, there has been growing interest and discussion in incorporating additional data attributes, such as textual data, into these systems (van der Aalst; Revina & Aksu, 2023).

Recent studies aim to improve existing process mining models by integrating new forms of data, including textual data along with digital logs. This integration is particularly relevant today, where machine learning techniques are widely used to derive valuable insights from textual information available in abundance on the internet, especially on social media platforms (Revina & Aksu, 2023).

1-1-          Review Methodology

The methodology of this literature review involved a multifaceted approach to identify and scrutinize relevant sources. This approach combined individual searches, references from reviewed articles, and the use of an AI tool. For the individual search, the online Mun library search engine was utilized, not only because of its reliability but also for its access to a wide range of authentic academic databases. Keywords such as “process mining,” “literature review,” and “review,” or any combination thereof, were primarily employed. This strategy yielded several pertinent articles that were reviewed for the purpose of this literature study. Furthermore, an initial reading of a few articles helped identify the most influential works by pioneers and renowned authors in the field, like van der Aalst, which were then explored further through a snowballing method. Within some of these articles, significant literature reviews were cited, which were also examined for this review. For instance, the review articles by Tiwari et al. (2008) and Emamjome et al. (2019) were referenced in multiple articles as key literature in the process mining research domain. Lastly, an AI tool named IRIS was employed to visually locate articles related to the topic. This tool operates by accepting an article title and mapping all related conceptual domains and their underlying articles, ensuring that all crucial articles related to process mining were reviewed. The article selected for mining the field of study was “Opportunities and Challenges for Process Mining in Organizations: Results of a Delphi Study” by Martin et al. (2021). This article was chosen because it was recently published, likely encompassing a comprehensive review of prior work. Additionally, notable figures in the field of process mining, such as van der Aalst, contributed to this study. The topic also indicates that the authors have given broad consideration to the future (Delphi). The graph below presents the results:

                                                                                          Fig1. Conceptual map of related research to the chosen article

By clicking on each segment, a deeper layer of related research is revealed. For instance, clicking on the mining sections displays various types of mining. Given our focus on process mining, we delve deeper into the process area, where we discover sixteen articles on related topics.

                                                                                             Fig2. The list of related articles based on the general topic

The initial idea for designing the literature review outline intuitively came to mind, encompassing theoretical backgrounds, issues and challenges, techniques and tools, process mining in practice, data requirements, and future research directions. Upon reviewing various articles, some of these sections and their subsections were revised.

1-2-          Literature review outline

This literature review offers a comprehensive examination of process mining. In the following section, we introduce some theoretical foundations, including definitions, types, perspectives, history, and methodologies of process mining. Subsequently, we will discuss the issues and challenges associated with process mining, which include process mining-specific issues, internal process mining issues, and external process mining issues. Then, we mention the techniques and tools of process mining, which are utilized for both process mining and the visualization of outputs. Next, the application of process mining in practice is discussed, followed by the data requirements for process mining. Finally, we will present some directions for future research.

2.    Theoretical Backgrounds

2.1.           Definitions (What is process mining?)

Process mining, originally referred to as workflow mining (Emamjome et al., 2019), encompasses a set of practices that apply data-mining techniques to process execution data to reconstruct actual business processes within organizations (Tiwari et al., 2008; Martin et al., 2021). It primarily utilizes real-life data, typically known as event logs, to discover, evaluate, or enhance business processes in organizations (De Medeiros et al., 2003). The definition of process mining has varied, at times being narrowly defined as a data-mining technique, and at other times broadly considered as a field of study comprising tools, techniques, and research methodologies (Emamjome et al., 2019). This data-driven approach seeks to uncover the realities of daily business operations rather than relying on perceptions of managers who design and implement processes based on theoretical models. The former approach is known as an evidence-based approach, while the latter is referred to as assumption-based process modeling. Both approaches undoubtedly aim to help organizations optimize their processes, identify bottlenecks, and improve overall efficiency. De Medeiros et al. (2003) highlight the differences between these two approaches, noting that process mining, or what they refer to as workflow mining, tends to be a bottom-up approach rather than a top-down approach that follows a pre-designed process model. Similarly, Zerbino et al. (2021) categorize two approaches to business process management, including the process model-driven approach and the data-driven approach, and consider process mining as a bridge that could fill the gap between these two legitimate approaches.

2.2.           Process mining types (Initiatives and aims)

Today’s business market and the new forms of competition among companies lead organizations to seek smarter ways to focus on their operations, and process mining has been recognized as one of these alternatives (Revina & Aksu, 2023). There are various types of process mining, each related to the aim of the project or, more precisely, the research question initially defined. According to the literature, there are at least three main reasons why organizations use process mining:

• For discovery (identifying new process models based on log data),
• For conformance (comparing existing models with log data to identify discrepancies),
• For enhancement (modifying existing process models based on insights gained from log data) (De Medeiros et al., 2003; Zerbino et al., 2021).

Clearly, these approaches are not mutually exclusive, meaning that one or all goals could be achieved in a process mining project, from discovering new or hidden workflows to comparing process modeling in theory and practice to enhancing business processes. Emamjome et al. (2019) discuss similar points, referring to five different types of process mining, including model discovery, conformance checking, performance checking, social network analysis, and comparative analysis. Process model discovery involves creating a process model from an event log without using any pre-existing process model. It aims to map out the actual processes as they are executed, providing visibility into the workflow. Conformance checking compares the observed event log (actual process) with an existing process model (expected process) to identify deviations, helping to understand whether the actual processes conform to the designed processes and where discrepancies lie. Performance checking focuses on analyzing the performance of processes, such as bottlenecks, service times, and throughput times, going beyond mere conformance to assess the efficiency of process execution. The social network analysis (SNA) approach examines the social dimensions of process mining, focusing on the interactions between resources (e.g., employees) involved in the process. It can reveal patterns of collaboration and communication flows within an organization. Finally, comparative analysis involves comparing different processes or the same process over different periods or under different conditions, useful for identifying best practices, trends, and areas for improvement by highlighting differences and similarities.

2.3.           Process mining perspectives

The types of process mining are primarily related to the objectives of conducting process mining, either conceptually or empirically (i.e., why do we need to conduct process mining?), while process perspectives deal with the angle of observation or scope of focus when analyzing process mining outcomes (i.e., what should we focus on while conducting process mining?). According to the definitions of business processes (Dumas et al., 2018), a process can be evaluated based on each of its components, namely the steps (activities), the timing of task execution or delays between activities, the individuals involved in performing the tasks, and the services or products resulting from the process. These perspectives are identified as the process, time, organizational, and case perspectives by Zerbino et al. (2021).

2.4.           History and roots

More than two decades after the advent of process mining, it is considered a mature body of knowledge (Emamjome et al., 2019). Process mining, a derivative of data mining—or what Emamjome et al. (2019) call a data-driven approach—primarily aims to extract meaning, knowledge, or wisdom from vast data sources (Tiwari et al., 2008). Agrawal et al. (1998) are widely recognized as pioneers in the field of process mining. They developed an algorithmic approach that enabled the construction of process flow graphs from execution logs generated by workflow applications. Cook and Wolf (1998) were among the first to attempt to discover software process models from event logs. Van der Aalst et al. (2004) further expanded process mining concepts by comparing various methods used to extract process models from data, ensuring that the outcomes of process mining techniques were congruent with typical workflow meta-models. Two meta-models have been employed in process mining techniques, including graph-oriented and block-oriented models. Some researchers, like Agrawal et al. (1998), considered net-based languages as a distinct meta-model, adding to the previously known models to encompass them all under the umbrella of a workflow language. Conversely, others, such as De Medeiros et al. (2003), argued that even net-based models like Petri nets are a type of graph-based meta-models. One common form of graph-oriented meta-model is the directed graph, which was among the techniques used by Agrawal et al. (1998) for process mining.

2.5.           Process mining methodology

Some research has been conducted to design a methodology for process mining. The process mining procedure usually begins with a problem or objective, as mentioned previously. The relevant techniques and tools are then selected based on the identified problems. Subsequently, data for the operation needs to be collected or gathered if it already exists. The data must be prepared for the operation, and by preparation, we mean addressing the main data issues, which will be discussed in subsequent sections. Following this, mining is conducted on the cleaned and integrated data, followed by the visualization of data, which is preferred to be in a graphical format. The results are then evaluated by stakeholders, and feedback is received to improve the process if possible. Finally, the confirmed results are utilized to implement the required changes. However, not every research project or process mining project strictly adheres to these steps. Moreover, the quality of each step can vary among research articles. This aspect, referred to as the thoroughness of a process mining technology, was measured by Emamjome et al. (2019) in their work.

3.    Issues and Challenges of process mining

Process mining issues can be categorized into three main areas. The first category stems from process mining itself and is referred to as process mining issues. The second group pertains to the internal environment of an organization and relates to the context in which process mining is conducted. The third category originates from the external environment, which encompasses broader areas surrounding process mining, including cultural, social, economic, political, environmental, and legal issues.

                                                                                                                     Fig3. Three types of PM issues

3.1.           External process mining issues

3.1.1.      Process mining and morality

The traditional approach to process modeling is primarily apparent to employees, as they are often instructed by their managers to implement the processes. Thus, their actions are confined to the scope of their responsibilities. However, in process mining, since the data is typically collected from various sources and information systems, not all employees are cognizant of such data exploitation. According to the human resource policies of some countries, the use of event logs is justified only if it cannot be traced back to its user or creator. Yet, in practice, logs can often be traced back to the user, which is where the question of morality arises (Chiu & Jans, 2019).

3.2.           Internal process mining issues

3.2.1.      Organizational issues

Mans et al. (2013) highlighted proper project management as an organizational issue in process mining (PM). Similarly, the lack of a governance mechanism and discrepancies between the design and implementation stages of PM projects were pointed out by Syed et al. (2020). Martin et al. (2021) discussed the challenges of aligning PM initiatives with strategy and operations, managing increased transparency through the use of PM, selecting the appropriate processes for PM application, and identifying sufficient Key Performance Indicators (KPIs) to quantitatively measure the value of PM as some of the organizational issues associated with PM.

3.2.2.      Human resource issues

Mans et al. (2013) mentioned the lack of domain expertise and the presence of technical PM skills as some of the soft issues related to PM. Similarly, Syed et al. (2020) also referred to the lack of PM training as a limitation.

3.3.           Process mining issues

3.3.1.      Data issues and preparation

One of the primary issues in process mining that directly impacts its quality and level of success is data quality. The data utilized in process mining is at risk of being low-quality, missing, or inefficient. The challenge lies in handling missing, incorrect, or poorly formatted data, given that data serves as the input for each process mining tool and technique. These efforts are so crucial that they are considered one of the main stages, known as data pre-processing, in process mining methodologies (Emamjome et al., 2019). Martin et al. (2021) highlight issues such as data access, the high cost of data preparation, and data quality as some of the data-related challenges in PM. Aalst and Weijters (2004) identify several data issues, including noise in logged data, which leads to inaccurate or incomplete process mining results. Managing heterogeneous results derived from information systems based on different platforms is another data issue mentioned by Tiwari et al. (2008).

3.3.2.      Process complexity issues:

According to Aalst and Weijters (2004), the presence of hidden tasks that are not captured in the data arises from the complexity of the process. Tiwari et al. (2008) also highlight several process mining issues, including the occurrence of duplicate tasks, non-free choice constructs, challenges in mining loops, and dealing with different perspectives. Among the most common problems in process mining identified in the articles reviewed by these authors were mining loops.

The graph below displays the frequency of process mining issues as identified in articles reviewed by Tiwari et al. (2008), covering the period from 1998 to 2008.

                                                                                       Fig 4 . The frequency of Process mining problems from 1998-2008 (Tiwari et al., 2008)

3.3.3.      Technical challenges

Both heuristic and soft computing techniques may present certain issues; that is, they are not universally foolproof solutions. It is argued that heuristic techniques often struggle to manage very complex process constructs. Conversely, soft computing techniques may fail to address some mining issues, such as noisy data, short loops, hidden activities, non-free choice constructs, and so forth. Additionally, these techniques may introduce new issues, such as permitting extra behavior not found in logs. The main aim of designing new techniques, such as genetic algorithms, is to tackle the issues mentioned previously (Tiwari et al., 2008).

3.3.4.      Tool’s issues

Like any data-driven approach, data mining presents implementation challenges (Van der Aalst et al., 2003). Furthermore, accessing a universally accepted data mining tool is difficult due to the lack of standardization and the significant number of business exceptions.

4.    Techniques (approaches) and tools (software)

Generally, tools and techniques in process mining are used for two main purposes: one is for mining data or processes, and the second is to present the outcome in various types of process models. This categorization stems from the methodology of process or data mining, which, in simple terms, involves first analyzing the data and then attempting to visualize it (Emamjome et al., 2019).

4.1.           Techniques to mining processes

Over the last few decades, researchers have adopted various approaches to process mining. Some process mining approaches are based on heuristic algorithms, also known as rules of thumb, which typically rely on certain assumptions about business process patterns. Tiwari et al. (2008) observed that nearly half of the articles they reviewed utilized a custom algorithmic approach to process mining instead of employing previously known methods. This trend may indicate either the limitations of the available methods or the rapidly growing field of process mining that encourages researchers to develop and expand their own approaches. However, over time, other approaches began to employ soft computing techniques such as genetic algorithms and neural network technologies, though these were not initially widely used by researchers.

Cook and Wolf (1998) introduced three statistical analysis methods that could be applied to mining tasks, although their original intention was not specifically for process mining. These methods include RNet, Markov, and Ktail. In the RNet approach, previous behavior is examined to identify the current state of a process. The Markov approach analyzes both past and future behavior to predict the current state, while Ktail similarly focuses on future behavior to describe the current state.

4.2.           Techniques to model (visualize) process

Several techniques are utilized to conduct process mining, including genetic algorithms, a general algorithmic approach, the Markovian approach, neural networks, and cluster analysis (Tiwari et al., 2008). Genetic algorithms operate based on the principles of Darwinian natural selection (De Medeiros et al., 2003). The general algorithmic approach is designed so that individual authors can apply it to process mining (van der Aalst and Weijters, 2004). As previously mentioned, the Markovian approach defines a potential current state by examining past and future behavior (Cook and Wolf, 1998). Neural networks function similarly to the human mind, learning from data by recognizing different data patterns (Cook and Wolf, 1998). Cluster analysis operates by dividing solutions into subgroups (Schimm, 2004).

Tiwari et al. (2008) documented the frequency of the aforementioned techniques used for either mining or modeling processes in articles from 1998 to 2008 (the time they wrote their article).

Fig 5. The frequency of Process mining techniques used from 1998-2008 (Tiwari et al., 2008)

Emamjome et al. (2019) introduced seven algorithms, including Fuzzy Miner, Heuristic Miner, Inductive Miner, Alpha Miner, Dotted Chart, Social Network Miner, and Clustering. Fuzzy Miner is one of the oldest yet simplest process mining techniques, offering flexibility that makes it suitable for complex and less structured processes. Its user-friendliness and capability to visualize complex processes are two main advantages. The Heuristic Miner technique strikes a good balance between accuracy and simplicity, with paths between activities connected based on heuristics. However, its use has declined as more powerful techniques have emerged that better address noise and complex behavior. The Inductive Miner, introduced in 2016, has proven effective at dealing with noise and missing or incomplete event logs. Unlike the Heuristic Miner, it is both sound and understandable. The Alpha Miner is one of the earliest process mining algorithms and is still used to some extent, despite limitations with loops or parallel executions. Social Network Analysis (SNA) is a prevalent application area in process mining, focusing on the relationships between process constructs such as people, departments, and resources. This technique can identify issues with communication, bottlenecks, and collaboration efficiency within organizations, extending somewhat beyond the flow of activities in process models.

Zerbino et al. (2021) categorized the process mining algorithms into three families: Fuzzy, Heuristic, and Alpha, ranked by their popularity (frequency of use in the articles reviewed in their study). They also noted that some algorithms, like the Inductive Miner, have been gaining popularity recently.

4.3.           Tools (Software)

Numerous software tools have been developed to facilitate process mining in research or business projects. InWolvE is one such tool, introduced by Herbst and Karagiannis (2004) for the first time. This software not only supported some of the algorithms available at the time but also was designed to accommodate four types of process models: injective sequential, injective parallel, non-injective sequential, and non-injective parallel. The distinction between injective and non-injective classes hinges on whether the nodes of a process are unique (not repeated in the process) or not. The classification of sequential versus parallel indicates whether a process diverges at its junctions (different scenarios) or not. The non-injective sequential class could be considered an inductive class to model a vast number of processes (Tiwari et al., 2008).

There are mainly two platforms, open-source and closed-source, on which many process mining (PM) tools are based. One of the most popular open-source platforms is ProM. On the commercial front, tools include, but are not limited to, Apromore, Celonis, and Disco (van der Aalst, 2016).

5.    Process mining in practice

Martin et al. (2021) categorize the process mining literature into business reports and academic research. Academic research is further divided into two main groups: one focusing on conceptual research, which includes conceptual frameworks, techniques, and algorithms, and the other on empirical research, which involves applying process mining for various purposes using event logs, often within a case study. Empirical studies can also be divided into two sub-groups based on whether they use real or simulated data, each with its own advantages and disadvantages (Zerbino et al., 2021).

As Emamjome et al. (2019) observed, the number of case studies related to process mining in practice has been increasing since 2014. This observation is corroborated by Zerbino et al. (2021), whose work indicates that more than two-thirds of the 145 reviewed articles focus on empirical research, in contrast to conceptual research, which comprises only one-third of the studies.

Process mining is predominantly used in healthcare and education. Although the authors identified seven common domains in their work, it appears they did not encompass other industries such as banking, trade, and the IT service and product sectors.

Fig6. The frequency of process mining articles in various domains (Emamjome et al. ,2019)

There may be several reasons why process mining is more prevalent in certain domains than in others. Zerbino et al. (2021) suggest that each domain may have distinct process characteristics, leading to operations that are either more complex or more straightforward. For example, healthcare often involves more process variants (scenarios) compared to guidelines or standard procedures, presenting challenges for process mining but also opportunities for improvement that could address social and economic concerns. In some domains, processes may be less complex, and consequently, there is less scope for improvement, resulting in lower interest in conducting process mining. Another reason for the popularity of healthcare, as mentioned by Zerbino et al. (2021), is the widespread use of human-computer tools (such as workflow systems) that generate sufficient event logs.

Thanks to Zerbino et al. (2021), the contribution of process mining research to business functions has been explored, with Operations being the most popular business function addressed by process mining efforts. Studies have also been conducted in infrastructure and service, with only a few in HRM, R&D, Marketing & Sales, and Logistics. It is evident that the most frequent managerial focus of empirical works is on operational aspects, followed by tactical, tactical-operational, and tactical-strategic focuses.

6.    Data requirements, challenges, and solutions for a having a successful process mining

In this section of the literature review, process mining is examined from the data perspective, which is essentially regarded as the raw material of process mining in organizations. A crucial concept here is the event log, recognized as one of the primary factors in the initial stages of process mining. Collecting and utilizing event logs in real-life situations is not straightforward and sometimes not even feasible (Zerbino et al., 2021). The greater the presence of the human factor in a process, the more challenging it becomes for person-to-person tools (P2P) to generate event logs, resulting in fewer opportunities for conducting process mining. Such data can be created and gathered more easily with person-to-application tools (P2A), such as workflow management systems, where both human and computer involvement exists (Zerbino et al., 2021). Therefore, even if it is possible to use event logs, retrieving data from data storages in the correct form for real-time use is a complex task. More specifically and technically, the challenge involves obtaining, transforming, organizing, and deriving data and process information from databases.

One of the primary data requirements for enabling process mining capability is data integration. This need arises from the business process definition, which extends the domain of a process beyond a single organizational division to broader organizational sections or business functions. In other words, business processes often span various organizational divisions, transforming an input into an output, typically a product or service. Consequently, it is clear that business processes utilize various data coming from different information systems or data sources. Similarly, process mining would require data from diverse sources, necessitating data integration (Aouachria et al., 2023).

The challenge of acquiring high-quality, comprehensive data from various sources within complex systems represents a significant barrier to effectively leveraging process mining (van der Aalst and Weijters, 2004). Addressing issues such as complex process structures, time discrepancies, noise, incomplete logs, and events of varying granularities requires robust techniques to maintain data integrity. Essential practices include data cleansing to minimize noise and anomalies, data authentication to ensure accuracy, and data validation to comply with established standards. These practices not only enhance data quality but also facilitate smoother data integration, potentially easing the challenges associated with integration in terms of time and resources.

7.    Future research direction

Since the advent of process mining, event logs have always been the cornerstone and are known as the heart of process mining. In today’s data-centric world, however, there is a significant investment in other types of data, particularly textual data. Despite the majority of companies’ data today consisting of textual data (around 80%), there have been limited efforts to incorporate it into process mining (Revina and Aksu, 2023).

One concerning trend in process mining case studies is that although the domain of process mining has been expanding, the thoroughness of these studies has decreased, described metaphorically by Emamjome et al. (2019) as “a shallow ocean.” This implies that process mining has not yet been able to deliver robust outcomes to real-world problems (Martin et al., 2021). They suggest developing guidelines so that the knowledge of mature process mining can be more effectively transferred from well-established researchers and experts in this field to newcomers and novice practitioners.

Moreover, addressing the complexity of new process mining techniques is a challenge that current advocates need to tackle (Imran et al., 2022).

Another fact is that only a minority of process mining projects have aimed for process enhancement, which is one of the most important objectives of business process management in organizations (Martin et al., 2021). Encouraging experts and research scholars to focus more on process enhancement in their process mining projects could unveil the true value of process mining to both academia and the business market.

In terms of business functions, PM could play a more pivotal role beyond Operations. Although there has been a recent focus on the service orientation of business activities, only a few process mining efforts have addressed service operations and activities. After-sales activities, which could significantly benefit from PM, are one area for potential future exploration (Zerbino et al., 2021).

PM could also significantly impact Logistics, including transportation, warehousing, and handling, where many event logs are created and recorded by devices such as workflow systems, mobile devices, and Internet of Things (IoT) sensors (Zerbino et al., 2021).

Sales & Marketing is another area that could benefit more from PM, especially in sales processes. The scarcity of process mining research in this area may be due to sales and marketing workflow systems typically not supporting event logs and the area being mined through other methods. However, applying process mining techniques from a customer experience perspective, or what could be termed a customer journey, might positively affect customer satisfaction and, consequently, organizational profits (Zerbino et al., 2021).

The relevance of PM in infrastructure activities, including Accounting, Finance, Quality Assurance, Document Management, and IT Management, as well as in procurement, stems from modern business approaches such as digitalization and paperless back-office operations.

HRM could also be positively impacted by PM, particularly in person-to-application sections such as payroll and time and attendance management (Zerbino et al., 2021).

Regarding managerial focus, most PM research has addressed operational and tactical management levels, with few studies focusing on the top management level of organizations. Although this may be due to the more person-to-person activities and the intuitive aspects of analysis at the top layer of organizations that cannot be recorded by event logs, some top management activities such as decision-making, corporate social responsibilities, or strategic and operational planning could be considered processes and, therefore, recorded by event logs in workflow systems (Martin et al., 2021).

As discussed in previous sections, the imbalanced distribution of research studies across various domains could trigger future research activities in process mining. Areas such as Farming, Energy & Materials, Consumer Goods, Consumer Services, Utilities, Financial, and Technologies could be untouched territories for future PM research and business projects (Zerbino et al., 2021).

References:

Agrawal, R., Gunopulos, D., & Leymann, F. (1998). Mining process models from workflow logs. ADVANCES IN DATABASE TECHNOLOGY – EDBT’98, 1377, 467–483. https://doi.org/10.1007/BFb0101003

Aouachria, M., Leshob, A., Ghomari, A. R., & Aouache, M. (2024). A Process Mining Method for Inter-organizational Business Process Integration. ACM Transactions on Management Information Systems, 15(1), 1–29. https://doi.org/10.1145/3638062

Chiu, T., & Jans, M. (2019). Process mining of event logs: A case study evaluating internal control effectiveness. Accounting Horizons, 33(3), 141–156. https://doi.org/10.2308/acch-52458

Cook, J., & Wolf, A. (1998). Discovering models of software processes from event-based data. ACM Transactions on Software Engineering and Methodology, 7(3), 215–249. https://doi.org/10.1145/287000.287001

De Medeiros, A. K. A., Van Der Aalst, W. M. P., & Weijters, A. J. M. M. (2003). Workflow Mining: Current Status and Future Directions. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 2888, 389–406. https://doi.org/10.1007/978-3-540-39964-3_25

Dumas, M. (2018). Fundamentals of Business Process Management (2nd ed. 2018.). Springer Nature. https://doi.org/10.1007/978-3-662-56509-4

Emamjome, F., Andrews, R., & ter Hofstede, A. H. M. (2019). A Case Study Lens on Process Mining in Practice. Lecture Notes in Computer Science (Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 11877, 127–145. https://doi.org/10.1007/978-3-030-33246-4_8

Herbst, J., & Karagiannis, D. (2004). Workflow mining with InWoLvE. Computers in Industry, 53(3), 245–264. https://doi.org/10.1016/j.compind.2003.10.002

Imran, M., Ismail, M. A., Hamid, S., & Nasir, M. H. N. M. (2022). Complex process modeling in Process mining: A systematic review. IEEE Access, 10, 1–1. https://doi.org/10.1109/ACCESS.2022.3208231

Mans, R., Reijers, H. A., Berends, H., Bandara, W., & Prince, R. (2013). Business process mining success. In Proceedings of the 2013 European Conference on Information Systems (p. 89). https://research.tue.nl/files/3943427/352067450737801.pdf

Martin, N., Fischer, D. A., Kerpedzhiev, G. D., Goel, K., Leemans, S. J. J., Röglinger, M., van der Aalst, W. M. P., Dumas, M., La Rosa, M., & Wynn, M. T. (2021). Opportunities and Challenges for Process Mining in Organizations: Results of a Delphi Study. Business & Information Systems Engineering, 63(5), 511–527. https://doi.org/10.1007/s12599-021-00720-0

Revina, A., & Aksu, Ü. (2023). An approach for analyzing business process execution complexity based on textual data and event log. Information Systems (Oxford), 114, 102184-. https://doi.org/10.1016/j.is.2023.102184

Schimm, G. (2003). Mining Most Specific Workflow Models from Event-Based Data. BUSINESS PROCESS MANAGEMENT, PROCEEDINGS, 2678, 25–40. https://doi.org/10.1007/3-540-44895-0_3

Syed, R., Leemans, S. J. J., Eden, R., & Buijs, J. A. C. M. (2020). Process Mining Adoption: A Technology Continuity Versus Discontinuity Perspective. In Business Process Management Forum (Vol. 392, pp. 229–245). Springer International Publishing. https://doi.org/10.1007/978-3-030-58638-6_14

Tiwari, A., Turner, C. J., & Majeed, B. (2008). A review of business process mining: state-of-the-art and future trends. Business Process Management Journal, 14(1), 5–22. https://doi.org/10.1108/14637150810849373

van der Aalst, W. M. P., van Dongen, B. F., Herbst, J., Maruster, L., Schimm, G., & Weijters, A. J. M. M. (2003). Workflow mining: A survey of issues and approaches. Data & Knowledge Engineering, 47(2), 237–267. https://doi.org/10.1016/S0169-023X(03)00066-1

van der Aalst, W. M. P., & Weijters, A. J. M. M. (2004). Process mining: a research agenda. Computers in Industry, 53(3), 231–244. https://doi.org/10.1016/j.compind.2003.10.001

van der Aalst, W. (2012). Process Mining: Overview and Opportunities. ACM Transactions on Management Information Systems, 3(2), 1–17. https://doi.org/10.1145/2229156.2229157

van der Aalst, W. M. P. (2016). Process Mining: Data Science in Action (2nd ed. 2016 edition.). Springer Berlin / Heidelberg. https://doi.org/10.1007/978-3-662-49851-4