This thesis explores the possibilities of probabilistic process modelling for the Computer Supported Cooperative Work (CSCW) systems in order to predict the behaviour of the users present in the CSCW system. Toward this objective applicability, advantages, limitations and challenges of probabilistic modelling are excavated in context of CSCW systems. Finally, as a primary goal seven models are created and examined to show the feasibilities of probabilistic process discovery and predictions of the users behaviour in CSCW systems.
The goal of this thesis is to create a recommender system (RS) for business processes, based on the existing ProM plugin RegPFA. To accomplish this task, firstly an interface must be created that sets up and expands a database receiving probabilistic finite automata (PFA) created by RegPFA in tsml format as input. Secondly, a Java program must be designed that uses said database to recommend the process elements that are most likely to follow a given sequence of process elements.
Business rules have become an important tool to warrant compliance at their business processes. But the collection of these business rules can have various conflicting elements. This can lead to a violation of the compliance to be achieved. This conflicting elements are therefore a kind of inconsistencies, or quasi incon- sistencies in the business rule base. The target for this thesis is to investigate how those quasi inconsistencies in business rules can be detected and analyzed. To this aim, we develop a comprehensive library which allows to apply results from the scientific field of inconsistency measurement to business rule formalisms that are actually used in practice.
Recently the workflow control as well as compliance analysis of the Enterprise Resource Planning systems are of a high demand. In this direction, this thesis presents the potential of developing a Workflow Management System upon a large Enterprise Resource Planning system by involving business rule extraction, business process discovery, design of the process, integration and compliance analysis of the system. Towards this, usability, limitations and challenges of every applied approach are deeply explained in the case of an existing system named SHD ECORO.
Business Process Querying (BPQ) is a discipline in the field of Business Process Man- agement which helps experts to understand existing process models and accelerates the development of new ones. Its queries can fetch and merge these models, answer questions regarding the underlying process, and conduct compliance checking in return. Many languages have been deployed in this discipline but two language types are dominant: Logic-based languages use temporal logic to verify models as finite state machines whereas graph-based languages use pattern matching to retrieve subgraphs of model graphs directly. This thesis aims to map the features of both language types to features of the other to identify strengths and weaknesses. Exemplarily, the features of Computational Tree Logic (CTL) and The Diagramed Modeling Language (DMQL) are mapped to one another. CTL explores the valid state space and thus is better for behavioral querying. Lacking certain structural features and counting mechanisms it is not appropriate to query structural properties. In contrast, DMQL issues structural queries and its patterns can reconstruct any CTL formula. However, they do not always achieve exactly the same semantic: Patterns treat conditional flow as sequential flow by ignoring its conditions. As a result, retrieved mappings are invalid process execution sequences, i.e. false positives, in certain scenarios. DMQL can be used for behavioral querying if these are absent or acceptable. In conclusion, both language types have strengths and are specialized for different BPQ use cases but in certain scenarios graph-based languages can be applied to both. Integrating the evaluation of conditions would remove the need for logic-based languages in BPQ completely.
To construct a business process model manually is a highly complex and error-prone task which takes a lot of time and deep insights into the organizational structure, its operations and business rules. To improve the output of business analysts dealing with this process, different techniques have been introduced by researchers to support them during construction with helpful recommendations. These supporting recommendation systems vary in their way of what to recommend in the first place as well as their calculations taking place under the hood to recommend the most fitting element to the user. After a broad introduction into the field of business process modeling and its basic recommendation structures, this work will take a closer look at diverse proposals and descriptions published in current literature regarding implementation strategies to effectively and efficiently assist modelers during their business process model creation. A critical analysis of presentations in the selected literature will point out strengths and weaknesses of their approaches, studies and descriptions of those. As a result, the final concept matrix in this work will give a precise and helpful overview about the key features and recommendation methods used and implemented in previous research studies to pinpoint an entry into future works without the downsides already spotted by fellow researchers.
The status of Business Process Management (BPM) recommender systems is not quite clear as research states. The use of recommenders familiarized itself with the world during the rise of technological evolution in the past decade.Ever since then, several BPM recommender systems came about. However, not a lot of research is conducted in this field. It is not well known to what broad are the technologies used and how are they used. Moreover, this master’s thesis aims at surveying the BPM recommender systems existing. Building on this, the recommendations come in different shapes. They can be positionbased where an element is to be placed at an element’s front, back or to autocomplete a missing link. On the other hand, Recommendations can be textual, to fill the labels of the elements. Furthermore, the literature review for BPM recommender systems took place under the guides of a literature review framework. The framework suggests 5stages of consecutive stages for this sake. The first stage is defining a scope for the research. Secondly, conceptualizing the topic by choosing key terms for literature research. After that in the third stage, comes the research stage.As for the fourth stage, it suggests choosing analysis features over which the literature is to be synthesized and compared. Finally, it recommends defining the research agenda to describe the reason for the literature review. By invoking the mentioned methodology, this master’s thesis surveyed 18 BPM recommender systems. It was found as a result of the survey that there
are not many different technologies for implementing the recommenders. It was also found that the majority of the recommenders suggest nodes that are yet to come in the model, which is called forward recommending. Also, one of the results of the survey indicated the scarce use of textual recommendations to BPM labels. Finally, 18 recommenders are considered less than excepted for a developing field therefore as a result, the survey found a shortage in the number of BPM recommender systems. The results indicate several shortages in several aspects in the field of BPM recommender systems. On this basis, this master’s thesis recommends the future work on it the results.
The main goal of this paper is to ascertain, if neural networks (especially LSTM) are helpful in predicting processes by making predictions as accurately as possible.
TensorFlow is the used framework in Python to build recurrent neural networks. Two networks are built, whereby one is used for training and the other one for prediction.
Used datasets contain several processes with several events each. With those processes, the network ist trained and afterwards, the parameters are saved. The network for prediction uses these parameters to make predictions.
The neural network is able to make clear predictions about subsequent events. Even branches can be predicted.
When developed further, integration in other programs is possible. It is recommended to use unique names for the events or to rename them.
Unkontrolliert gewachsene Software-Architekturen zeichnen sich i.d.R. durch fehlende oder schlecht nachvollziehbare Strukturen aus. Hierfür können als Gründe beispielsweise mangelhafte Definitionen oder ein langsames Erodieren sein. Dies ist auch unter dem Begriff "Big Ball of Mud" bekannt. Langfristig erhöhen solche architekturellen Mängel nicht nur die Entwicklungskosten, sondern können letztendlich auch Veränderungen vollständig verhindern.
Die Software-Architektur benötigt somit eine kontinuierliche Weiterentwicklung, um solchen Effekten entgegen wirken zu können. Eine gute Software-Architektur unterstützt die Software-Entwicklung und erhöht die Produktivität. Auf der Ebene von Quellcode existieren bereits etablierte Vorgehensweisen zur kontrollierten Verbesserung der Qualität. Im Gegensatz hierzu existieren für Verbesserungen einer Software-Architektur jedoch keine allgemeingültigen Vorgehensweisen, welche unabhängig vom Anwendungsfall angewandt werden können. An diesem Punkt setzt die vorliegende Arbeit an.
Bisherige Arbeiten beschäftigen sich einerseits nur mit Teilpunkten des Problems. Anderseits existieren zwar bereits Vorgehensweisen zum Treffen von Architekturentscheidungen, jedoch agieren diese auf einer stark abstrakten Ebene ohne praktische Beispiele. Diese Arbeit stellt eine leichtgewichtige Vorgehensweise zum gezielten Verbessern einer Software-Architektur vor. Die Vorgehensweise basiert auf einem generischen Problemlösungsprozess. Auf dieser Basis ist ein Prozess zum Lösen von Problemen einer Software-Architektur entwickelt worden. Im Fokus der Arbeit stehen zur Eingrenzung des Umfanges architektonische Probleme aufgrund geforderter Variabilität sowie externer Abhängigkeiten.
Die wissenschaftliche Methodik, welcher der Arbeit zugrunde liegt, agiert im Rahmen der Design Science Research (DSR). Über mehrere Iterationen hinweg wurde eine Vorgehensweise entwickelt, welche sich an Softwareentwickler mit zwei bis drei Jahren Erfahrung und Kenntnissen über Grundlage der Softwareentwicklung und Software-Architektur richtet. Fünf Schritte inkl. Verweise auf aussagekräftige Literatur leiten Anwender anschließend durch den Prozess zur gezielten Verbesserung einer Software-Architektur.
Nowadays, almost any IT system involves personal data processing. In
such systems, many privacy risks arise when privacy concerns are not
properly addressed from the early phases of the system design. The
General Data Protection Regulation (GDPR) prescribes the Privacy by
Design (PbD) principle. As its core, PbD obliges protecting personal
data from the onset of the system development, by effectively
integrating appropriate privacy controls into the design. To
operationalize the concept of PbD, a set of challenges emerges: First, we need a basis to define privacy concerns. Without such a basis, we are not able to verify whether personal data processing is authorized. Second, we need to identify where precisely in a system, the controls have to be applied. This calls for system analysis concerning privacy concerns. Third, with a view to selecting and integrating appropriate controls, based on the results of system analysis, a mechanism to identify the privacy risks is required. Mitigating privacy risks is at the core of the PbD principle. Fourth, choosing and integrating appropriate controls into a system are complex tasks that besides risks, have to consider potential interrelations among privacy controls and the costs of the controls.
This thesis introduces a model-based privacy by design methodology to handle the above challenges. Our methodology relies on a precise definition of privacy concerns and comprises three sub-methodologies: model-based privacy analysis, modelbased privacy impact assessment and privacy-enhanced system design modeling. First, we introduce a definition of privacy preferences, which provides a basis to specify privacy concerns and to verify whether personal data processing is authorized. Second, we present a model-based methodology to analyze a system model. The results of this analysis denote a set of privacy design violations. Third, taking into account the results of privacy analysis, we introduce a model-based privacy impact assessment methodology to identify concrete privacy risks in a system model. Fourth, concerning the risks, and taking into account the interrelations and the costs of the controls, we propose a methodology to select appropriate controls and integrate them into a system design. Using various practical case studies, we evaluate our concepts, showing a promising outlook on the applicability of our methodology in real-world settings.
