工作流程（workflow）是商業流程自動化的一部份。一個工作流程是由完成一件工作所有可能執行的活動（activity）以及活動間在執行時的前後關係所構成。而工作流程的設計或改進舊有的工作流程是商業上很重要的工作，因為工作流程的好與壞會影響企業的競爭力。工作流程探勘（workflow mining）是利用資料探勘的技術，分析工作流程執行時所留下的流程執行記錄，還原出一個能夠產生這些記錄的工作流程模型（workflow model），而這個工作流程模型可做為設計新模型或改進既有模型的參考。 本研究針對我們所定義的工作流程模型，以一個未知的工作流程模型所產生的流程執行記錄（workflow log）當做輸入資料（input data），提出方法利用輸入資料還原一個能夠產生輸入資料中所有資料工作流程模型，且希望這個工作流程模型能與產生流程執行記錄之未知模型越相似越好。我們提出兩個還原工作流程模型的演算法，並利用precision和recall來評估還原的模型與未知模型間的相似程度，驗證我們所提出方法的效果。實驗結果顯示，我們的方法所還原的工作流程模型precision和recall值都能達到80%以上。 / The workflow plays an important role in business process automation. A workflow is composed of activities and causal relations between activities to complete a task. Workflow design and refinement are important tasks in business process reengineering. As a workflow is executed, the orders of the executed activities are recorded in workflow logs. Workflow mining utilizes the technology of data mining to analyze these workflow logs, and reconstruct a workflow model. In this thesis, we investigate the workflow mining problem to reconstrcuct the workflow model. Two algorithms are proposed to reconstruct a workflow model. We evaluate our proposed algorithms by precision and recall to measure the similarity between the constructed and the groundtruth models. The result of the experiment shows that our proposed methods can achieve 80% precision and 80% recall for the reconstruction of workflow models.

Contemporary workflow management systems are driven by explicit process models, i.e., a completely specified workflow design is required in order to enact a given workflow process. Creating a workflow design is a complicated time-consuming process and typically there are discrepancies between the actual workflow processes and the processes as perceived by the management. Therefore, new techniques for discovering workflow models have been required. Starting point for such techniques are so-called &ldquo / workflow logs&quot / containing information about the workflow process as it is actually being executed. In this thesis, new mining technique based on time information is proposed. It is assumed that events in workflow logs bear timestamps. This information is used in to determine task orders and control flows between tasks. With this new algorithm, basic workflow structures, sequential, parallel, alternative and iterative (i.e., loops) routing, and advance workflow structure or-join can be mined. While mining the workflow structures, this algorithm also handles the noise problem.

Many software companies today collect command usage data by monitoring and logging user interactions within their applications. This is not always utilised to its full potential, but with the use of state-of-the-art process mining techniques, this command usage log data can be used to gain insights about the users' workflows. These insights can then be used to improve the software application and boost user productivity and efficiency. One area where this might be especially relevant is within the radiology domain, where the radiology labour shortage renders every efficiency-improvement valuable. Connected to this, this thesis aimed to evaluate a number of process mining techniques on real radiology PACS command usage data from Sectra to identify which techniques might be useful for analysing user workflows. Three process discovery algorithms (Alpha, Heuristics, and Inductive Miner - infrequent) were used on two datasets and evaluated based on a number of quantitative metrics (fitness and simplicity) and qualitative aspects (interpretability and usefulness). The qualitative aspects of the resulting process models were assessed through interviews with domain experts at Sectra, and the Heuristics Miner was found to discover the most useful models that could be interpreted and analysed by domain experts, mainly due to its simpler process model notation. To reduce model complexity, three different filtering methods based on sequential pattern mining were evaluated as a pre-processing step before applying the discovery algorithms. This resulted in improvements for the Alpha and Inductive Miner - infrequent, although none of the methods improved the Heuristic Miner models against the baseline. Trace clustering was also explored to address model complexity with the aim of identifying trace execution variants. Several configurations of trace representation techniques and clustering algorithms were used, and the neural-network-based approaches, Word2Vec and Autoencoder, emerged as the alternatives that achieved the best scores in the clustering evaluation. A few clusters with well-separated trace execution variants were identified - although most clusters were still complex and dominated by similar events. Finally, a prototype application with integrated process mining concepts was created based on the findings from the previous interviews. This was then evaluated with domain experts at Sectra, with the aim of investigating what concepts are practically useful for assisting with domain analysis. The findings indicate a clear use case for such an application to analyse sequential relations and command usage patterns in PACS user workflows, providing a data-driven and on-demand approach for hypotheses testing. Simpler concepts like manual filtering and aggregation were found to be practically useful and prioritised by the domain experts, while the opinion was more divided on the automatic pre-processing methods.

abstract: The following is a case study composed of three workflow investigations at the open source software development (OSSD) based Apache Software Foundation (Apache). I start with an examination of the workload inequality within the Apache, particularly with regard to requirements writing. I established that the stronger a participant's experience indicators are, the more likely they are to propose a requirement that is not a defect and the more likely the requirement is eventually implemented. Requirements at Apache are divided into work tickets (tickets). In our second investigation, I reported many insights into the distribution patterns of these tickets. The participants that create the tickets often had the best track records for determining who should participate in that ticket. Tickets that were at one point volunteered for (self-assigned) had a lower incident of neglect but in some cases were also associated with severe delay. When a participant claims a ticket but postpones the work involved, these tickets exist without a solution for five to ten times as long, depending on the circumstances. I make recommendations that may reduce the incidence of tickets that are claimed but not implemented in a timely manner. After giving an in-depth explanation of how I obtained this data set through web crawlers, I describe the pattern mining platform I developed to make my data mining efforts highly scalable and repeatable. Lastly, I used process mining techniques to show that workflow patterns vary greatly within teams at Apache. I investigated a variety of process choices and how they might be influencing the outcomes of OSSD projects. I report a moderately negative association between how often a team updates the specifics of a requirement and how often requirements are completed. I also verified that the prevalence of volunteerism indicators is positively associated with work completion but what was surprising is that this correlation is stronger if I exclude the very large projects. I suggest the largest projects at Apache may benefit from some level of traditional delegation in addition to the phenomenon of volunteerism that OSSD is normally associated with. / Dissertation/Thesis / Doctoral Dissertation Industrial Engineering 2017

Mineração de processos é uma área de pesquisa relativamente recente que se situa entre mineração de dados e aprendizado de máquina, de um lado, e modelagem e análise de processos de negócio, de outro lado. Mineração de processos visa descobrir, monitorar e aprimorar processos de negócio reais por meio da extração de conhecimento a partir de logs de eventos disponíveis em sistemas de informação orientados a processos. O principal objetivo deste trabalho foi avaliar o contexto de aplicação de técnicas provenientes das áreas de inteligência computacional e de aprendizado de máquina, incluindo redes neurais artificiais. Para fins de simplificação, denominadas no restante deste texto apenas como ``redes neurais\'\'. e máquinas de vetores de suporte, no contexto de mineração de processos. Considerando que essas técnicas são, atualmente, as mais aplicadas em tarefas de mineração de dados, seria esperado que elas também estivessem sendo majoritariamente aplicadas em mineração de processos, o que não tinha sido demonstrado na literatura recente e foi confirmado por este trabalho. Buscou-se compreender o amplo cenário envolvido na área de mineração de processos, incluindo as principais caraterísticas que têm sido encontradas ao longo dos últimos dez anos em termos de: tipos de mineração de processos, tarefas de mineração de dados usadas, e técnicas usadas para resolver tais tarefas. O principal enfoque do trabalho foi identificar se as técnicas de inteligência computacional e de aprendizado de máquina realmente não estavam sendo amplamente usadas em mineração de processos, ao mesmo tempo que se buscou identificar os principais motivos para esse fenômeno. Isso foi realizado por meio de um estudo geral da área, que seguiu rigor científico e sistemático, seguido pela validação das lições aprendidas por meio de um exemplo de aplicação. Este estudo considera vários enfoques para delimitar a área: por um lado, as abordagens, técnicas, tarefas de mineração e ferramentas comumente mais usadas; e, por outro lado, veículos de publicação, universidades e pesquisadores interessados no desenvolvimento da área. Os resultados apresentam que 81% das publicações atuais seguem as abordagens tradicionais em mineração de dados. O tipo de mineração de processos com mais estudo é Descoberta 71% dos estudos primários. Os resultados deste trabalho são valiosos para profissionais e pesquisadores envolvidos no tema, e representam um grande aporte para a área / Mining process is a relatively new research area that lies between data mining and machine learning, on one hand, and business process modeling and analysis, on the other hand. Mining process aims at discovering, monitoring and improving business processes by extracting real knowledge from event logs available in process-oriented information systems. The main objective of this master\'s project was to assess the application of computational intelligence and machine learning techniques, including, for example, neural networks and support vector machines, in process mining. Since these techniques are currently widely applied in data mining tasks, it would be expected that they were also widely applied to the process mining context, which has been not evidenced in recent literature and confirmed by this work. We sought to understand the broad scenario involved in the process mining area, including the main features that have been found over the last ten years in terms of: types of process mining, data mining tasks used, and techniques applied to solving such tasks. The main focus of the study was to identify whether the computational intelligence and machine learning techniques were indeed not being widely used in process mining whereas we sought to identify the main reasons for this phenomenon. This was accomplished through a general study area, which followed scientific and systematic rigor, followed by validation of the lessons learned through an application example. This study considers various approaches to delimit the area: on the one hand, approaches, techniques, mining tasks and more commonly used tools; and, on the other hand, the publication vehicles, universities and researchers interested in the development area. The results show that 81% of current publications follow traditional approaches to data mining. The type of mining processes more study is Discovery 71% of the primary studies. These results are valuable for practitioners and researchers involved in the issue, and represent a major contribution to the area

Mineração de processos é uma área de pesquisa relativamente recente que se situa entre mineração de dados e aprendizado de máquina, de um lado, e modelagem e análise de processos de negócio, de outro lado. Mineração de processos visa descobrir, monitorar e aprimorar processos de negócio reais por meio da extração de conhecimento a partir de logs de eventos disponíveis em sistemas de informação orientados a processos. O principal objetivo deste trabalho foi avaliar o contexto de aplicação de técnicas provenientes das áreas de inteligência computacional e de aprendizado de máquina, incluindo redes neurais artificiais. Para fins de simplificação, denominadas no restante deste texto apenas como ``redes neurais\'\'. e máquinas de vetores de suporte, no contexto de mineração de processos. Considerando que essas técnicas são, atualmente, as mais aplicadas em tarefas de mineração de dados, seria esperado que elas também estivessem sendo majoritariamente aplicadas em mineração de processos, o que não tinha sido demonstrado na literatura recente e foi confirmado por este trabalho. Buscou-se compreender o amplo cenário envolvido na área de mineração de processos, incluindo as principais caraterísticas que têm sido encontradas ao longo dos últimos dez anos em termos de: tipos de mineração de processos, tarefas de mineração de dados usadas, e técnicas usadas para resolver tais tarefas. O principal enfoque do trabalho foi identificar se as técnicas de inteligência computacional e de aprendizado de máquina realmente não estavam sendo amplamente usadas em mineração de processos, ao mesmo tempo que se buscou identificar os principais motivos para esse fenômeno. Isso foi realizado por meio de um estudo geral da área, que seguiu rigor científico e sistemático, seguido pela validação das lições aprendidas por meio de um exemplo de aplicação. Este estudo considera vários enfoques para delimitar a área: por um lado, as abordagens, técnicas, tarefas de mineração e ferramentas comumente mais usadas; e, por outro lado, veículos de publicação, universidades e pesquisadores interessados no desenvolvimento da área. Os resultados apresentam que 81% das publicações atuais seguem as abordagens tradicionais em mineração de dados. O tipo de mineração de processos com mais estudo é Descoberta 71% dos estudos primários. Os resultados deste trabalho são valiosos para profissionais e pesquisadores envolvidos no tema, e representam um grande aporte para a área / Mining process is a relatively new research area that lies between data mining and machine learning, on one hand, and business process modeling and analysis, on the other hand. Mining process aims at discovering, monitoring and improving business processes by extracting real knowledge from event logs available in process-oriented information systems. The main objective of this master\'s project was to assess the application of computational intelligence and machine learning techniques, including, for example, neural networks and support vector machines, in process mining. Since these techniques are currently widely applied in data mining tasks, it would be expected that they were also widely applied to the process mining context, which has been not evidenced in recent literature and confirmed by this work. We sought to understand the broad scenario involved in the process mining area, including the main features that have been found over the last ten years in terms of: types of process mining, data mining tasks used, and techniques applied to solving such tasks. The main focus of the study was to identify whether the computational intelligence and machine learning techniques were indeed not being widely used in process mining whereas we sought to identify the main reasons for this phenomenon. This was accomplished through a general study area, which followed scientific and systematic rigor, followed by validation of the lessons learned through an application example. This study considers various approaches to delimit the area: on the one hand, approaches, techniques, mining tasks and more commonly used tools; and, on the other hand, the publication vehicles, universities and researchers interested in the development area. The results show that 81% of current publications follow traditional approaches to data mining. The type of mining processes more study is Discovery 71% of the primary studies. These results are valuable for practitioners and researchers involved in the issue, and represent a major contribution to the area

Describir los mejores procesos para ejecutar correctamente una estrategia de una forma eficiente y con calidad no es siempre una tarea fácil. La estandarizaci ón de procesos en general y de Vías Clínicas en particular requiere de potentes herramientas de especificación e implementación que apoyen a los expertos en diseño. La utilización de modelos de Flujos de Trabajo (del inglés Work ows) facilita a los expertos en diseño la creación las reglas de ejecución de sus sistemas como si fueran programadores. Aún así debido a la gran mutabilidad de los procesos reales, es muy difícil conocer como los procesos se están ejecutando en la realidad. La utilización de técnicas de reconocimiento de formas pueden ayudar a los expertos en procesos a inferir, a partir de muestras de ejecución pasadas, modelos que expliquen la forma en la que estos procesos están efectivamente ejecutándose. Este paradigma es conocido como Aprendizaje de Flujos de Trabajo (del inglés Work ow Mining). Los cambios de estado en los procesos de cuidado existentes en las Vías Clínicas se basan en los resultados de las acciones. Los modelos actuales de Aprendizaje de Flujos de Trabajo no recogen esta información en sus corpus. Por eso, los actuales sistemas de aprendizaje no cubren las necesidades de problemas complejos como es el caso de las Vías Clínicas. En esta Tesis se van a estudiar los modelos de representación, interpretación y aprendizaje de Flujos de Trabajo con la intención de proponer un modelo adecuado para resolver los problemas que impiden a los diseñadores de procesos complejos, como Vías Clínicas, utilizar técnicas de Aprendizaje de Flujos de Trabajo. Para ello se va a definir un nuevo paradigma adecuado para el apoyo al diseño de Vías Clínicas, además de proporcionar herramientas para su uso. Por esto en esta Tesis se presenta además un modelo de representación de Flujos de Trabajo con una alta expresividad y legibilidad, una herramienta software capaz de ejecutar y simular Flujos de Trab / Fernández Llatas, C. (2009). Representación, interpretación y aprendizaje de flujos de trabajo basado en actividades para la estandarización de vías clínicas [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/4562