Die Professur BKS bietet Abschlussarbeiten allgemein im Bereich Business Process Management an. Folgende Themen sind im Moment möglich:

  • Transformation of DMN-Tables into FCL-Rules (Implementation) [MSc]
    Next to business process models, which capture the procedural character of business processes, DMN-Tables can be used to model a declarative view on said processes. This thesis should aim at transforming DMN-tables into FCL-rules (a logic formalism). 
  • Mining (FCL-)Rules from Text (Implementation) [BSc/MSc]
    Having business rules stored in logical form poses many advantages for companies. However, there are often experts missing to initially create such logical expressions. This thesis should aim at implementing a miner, capable of parsing some form of textual rules into a logic program rule base. 
    Material: Bajwa, I. S., Lee, M. G., & Bordbar, B. (2011, March). SBVR Business Rules Generation from Natural Language Specification. In AAAI spring symposium: AI for business agility (pp. 2-8).
  • Survey of Compliance-Violation-Pattern Classifications [BSc/MSc]
    Many different "categories" of compliance violation patterns have been proposed. This thesis should aim at an extensive review and classification of proposed families of patterns.
    Material: Ramezani, E., Fahland, D., & van der Aalst, W. M. (2012, September). Where did i misbehave? diagnostic information in compliance checking. In International conference on business process management (pp. 262-278). Springer Berlin Heidelberg.
  • Mapping possibilities of graph-based and logic-based model query languages [MSc]
    Model query languages are used to find sections in conceptual models that comply with a specific structure or behavior. For instance, in Business Process Management endeavors, we can use those query languages to find weaknesses, inefficiencies or law violations in process models so we can improve them subsequently. Some query languages come from the field of logic, some others come from the field of graph theory. Both have their advantages and disadvantages. In order to exploit the advantages from both fields and develop query languages that include these advantages, we must analyze which capabilities of logic-based languages are already reproducible by graph-based query languages and vice-versa. The latter is the topic of the thesis.
  • Survey of Machine Learning approaches suitable for predictive process mining [BSc/MSc]
    Predictive process mining means to generate a process model out of log files of business software and use the generated model to predict the future behavior of currently running process instances. This can be used, for instance, to predict fraud, failure or other undesired behavior in business processes in order to intervene in time or to predict congestion in production lines. Predictive process mining approaches are rare, and those that already exist rely on a restricted set of predictive techniques from statistics. With the thesis, we would like to analyze the spectrum of available predictive techniques that could be suitable (and, probably, better than the currently used ones) for predictive process mining.
  • Survey of Business-Rule Vendors and Services [BSc/MSc]
    Many vendors and services that provide (international) rule bases have emerged, e.g. rule bases containing relevant laws that banks have to adhere to. This thesis should aim at an extensive review, comparison and description of available vendors and services.
  • Classification of Model Merging approaches [BSc/MSc]
    In  cases of company mergers or acquisitions, respective business processes of companies might have to be merged as well, desirably automatedly. Many merging approaches have been proposed. This thesis should aim at an extensive review and comparison of model merging approaches.
  • Survey of quantitative measures of compliance in process models [BSc/MSc]
    Detecting compliance of business processes is a central field of interest. Here, many approaches have been proposed towards finding the presence of errors in process models (e.g. the process as it is modelled violates a law). There has been recent attention towards analyzing the detected errors and comparing them. The intuition behind this is, that different errors may have a different "degree" of violating compliance, i.e. some errors can be said to be "worse/more significant" than others. This thesis should be aimed at an extensive review of proposed quantitative measures, capable of quantitatively assessing the degree/severity of compliance violations.
  • Survey of Ontology-Based Process Modeling [MSc]
    There has been recent attention directed towards extending classical business process models with semantic annotations, in order to generate a so-called ontology-based process model. Many proposals as to what benefits this artifacts can bring to companies have been proposed. This thesis should aim at an extensive review of literature on ontology-based process modeling.
  • Survey of Process Modeling Recommender Systems [BSc/MSc]
    Recommender Systems help process modelers in creating process models more quickly and accurately. Based on the lastly modeled elements, such systems provide suggestions what to model next or point to possible semantic flaws. These systems rely on different methodologies, e.g., ontologies or frequencies of already modeled elements in other models. The thesis should provide an overview of existing recommender systems approaches and the basic formal concepts they rely on.
  • Diverse Themen zur Erweiterung eines Modelierungstools in C# [BSc/MSc]
    Weiterentwicklung eines bestehenden Modelierungstools.  Je nach Absprache mit Herrn Delfmann.