Institute for Enterprise Systems

The success of information systems (IS) development strongly depends on the accuracy of the requirements gathered from users and other stakeholders. When developing a new IS, about 80 percent of these requirements are recorded in informal requirements documents (e.g., interview transcripts or discussion forums) using natural language. However, processing the resultant natural language requirements resources is inherently complex and often error prone due to ambiguity, inconsistency, and incompleteness. Thus, even highly qualified requirements engineers often struggle to process large amounts of natural language requirements resources efficiently and effectively. In this paper, we propose a design theory for requirement mining systems (RMSs) based on two design principles: (1) semi-automatic requirement mining and (2) usage of imported and retrieved knowledge. As part of an extensive design project, which led to these principles, we also implemented a prototype based on this design theory (REMINER). It supports requirements engineers in identifying and classifying requirements documented in natural language and allows us to evaluate the artifact’s viability and the conceptual soundness of our design. The results of our evaluation suggest that an RMS based on our proposed design principles can significantly improve recall while maintaining precision levels.

Abstract Agile methods are widely used in the software industry as a way to more rapidly develop and deliver new software. They define iterative work processes, advocate self‐organization and openness for change, and prescribe how software developers interact with each other and external stakeholders. Despite their popularity, it is unclear how agile methods influence work exhaustion in software developers and how developer skills play into this effect. On the one hand, agile methods may reduce software developers' work exhaustion by levelling out their workload across the entire duration of a project. On the other hand, agile methods exert a high level of pressure on software developers to continuously deliver working software, create many intensive social interactions, and to frequently adapt to changes. In light of these effects, prior research could not explain why some software developers become less exhausted from using agile methods, whereas others perceive the exact opposite. Based on the job demand‐control model, we develop a theoretical model connecting agile method use to individual developer skills and to two established determinants of employee exhaustion: role conflict and role ambiguity. We tested our research model in a field study among 1894 software developers in 217 project teams that used agile methods. The random coefficient modelling results show that agile method use facilitates the achievement of clear and unambiguous role perceptions and thereby reduces work exhaustion in developers, particularly if developers possess the organizational skills to effectively interact with others in their organization. We highlight implications for theory on the individual‐level effects of software development methods and provide practical insights for software companies.

The widespread use of lean and agile development methods shows a fundamental shift in how organizations try to cope with complexity and volatility issues. In large-scale settings, the coordination of many people often results in a team of teams setup. We introduce the multiteam systems perspective to describe different conceptual strategy types for inter-team coordination. These types are illustrated with examples from a large enterprise software development organization.

The goal of process discovery is to visualize event log data as a process model. In reality, however, these models are often highly complex. Process trace clustering is a well-studied and powerful technique to address this. It groups an event log into more cohesive sub logs, such that the discovered process models become less complex and easier to understand. Over the past 15 years, researchers proposed various approaches for trace clustering in process discovery. The developed approaches vary greatly with regard to algorithmic capacities, data characteristics, computational complexity, and integration of additional information. In this paper, we provide a state-of-the-art analysis of trace clustering by a) performing a systematic literature review, and b) proposing a generic framework for trace clustering. Eventually, our goal is to provide an overview of current trace clustering research and a basis for developing new methods and approaches to trace clustering.

The automatic, sensor-based assessment of human activities is highly relevant for production and logistics, to optimise the economics and ergonomics of these processes. One challenge for accurate activity recognition in these domains is the context-dependence of activities: Similar movements can correspond to different activities, depending on, e.g., the object handled or the location of the subject. In this paper, we propose to explicitly make use of such context information in an activity recognition model. Our first contribution is a publicly available, semantically annotated motion capturing dataset of subjects performing order picking and packaging activities, where context information is recorded explicitly. The second contribution is an activity recognition model that integrates movement data and context information. We empirically show that by using context information, activity recognition performance increases substantially. Additionally, we analyse which of the pieces of context information is most relevant for activity recognition. The insights provided by this paper can help others to design appropriate sensor set-ups in real warehouses for time management.

Although several theories could be applied to stimulate mobile health (mHealth) use and change people’s health behaviors, theory-driven mHealth designs are scarce. Instead, developers and researchers often implement behavior change techniques and mix multiple techniques without explicitly considering their underlying theoretical mechanisms. We call this practice into question and propose that combining behavior change techniques does not necessarily result in synergistic effects. Drawing on theories of protection motivation and social upward comparison, we engage in explanatory design theorizing to understand interactions of behavior change techniques and their implications for mHealth design. We design, theorize about, and evaluate four mHealth prototypes for stress alleviation. In a five-week field experiment with n = 138 participants and a subsequent qualitative substantiation, we show that there is a negative interaction effect of protection motivation and social upward comparison, rendering their combined application less effective. Our findings elicit mutual boundary conditions for theories of protection motivation and social upward comparison. If mechanisms of one theory are present, they restrict the effectiveness of mechanisms of the other theory. Thus, mHealth developers need to use caution when combining different behavior change techniques within one mHealth artifact.

The explosive growth of mobile devices and their widespread acceptance by customers along with the potential benefits of autoID technologies have prompted retailers to consider adoption of emerging technologies. Their motives are to enhance in-store customer shopping experience and to acquire an advantage in the competitive retail environment. Two interrelated issues nevertheless have been a hindrance: mobile shopping application usability and customers’ privacy concerns. Drawing on a recently developed conceptualisation of mobile application usability and the multidimensional developmental theory of privacy, we tackle these two issues. We theorise about the impact of artifact design on mobile application usability and the consequent impact of usability on customers’ privacy concerns and shopping efficiency. We tested our hypotheses in two retail store laboratory studies in which the participants were assigned to two different shopping tasks: general browsing (n = 194) and goal-directed shopping (n = 190). We found that adhering to mobile application usability principles could mitigate privacy concerns and consequently, improve shopping efficiency. Our findings suggest new avenues to alleviate customers’ privacy concerns using artifact design, thus complementing conventional approaches that focus on preventive measures to deal with the issue of privacy concerns.

Abstract Rapid GUI prototyping has evolved into a widely applied technique in early stages of software development to facilitate the clarification and refinement of requirements. Especially high-fidelity GUI prototyping has shown to enable productive discussions with customers and mitigate potential misunderstandings, however, the benefits of applying high-fidelity GUI prototypes are accompanied by the disadvantage of being expensive and time-consuming in development and requiring experience to create. In this work, we show RaWi , a data-driven GUI prototyping approach that effectively retrieves GUIs for reuse from a large-scale semi-automatically created GUI repository for mobile apps on the basis of Natural Language (NL) searches to facilitate GUI prototyping and improve its productivity by leveraging the vast GUI prototyping knowledge embodied in the repository. Retrieved GUIs can directly be reused and adapted in the graphical editor of RaWi . Moreover, we present a comprehensive evaluation methodology to enable (i) the systematic evaluation of NL-based GUI ranking methods through a novel high-quality gold standard and conduct an in-depth evaluation of traditional IR and state-of-the-art BERT-based models for GUI ranking, and (ii) the assessment of GUI prototyping productivity accompanied by an extensive user study in a practical GUI prototyping environment.

This record contains a full paper presented at the 5th Conference on Language Technologies (JT-2006), held in Ljubljana, Slovenia, in October 2006.

Functional and nonfunctional characteristics of software systems are defined by their architecture. Therefore, research streams such as Internet-of-Things (IoT) or component-based software engineering provide researchers and practitioners with construction guidelines for selected architectural characteristics. Current systems can be categorized in delivering services to the user and being engineered in a smart way. For example, services being provided by IoT-Systems must fulfill users’ goals in a highly dynamic and ad-hoc way. Consequently, this survey aims at answering various research questions regarding the methodical composition of system components and services. Furthermore, new research opportunities are sketched that should be tackled to make the scientific progress available to practitioners. Based on a systematic literature review from a software architecture point of view, in this paper we identify 75 primary studies for domain-specific IoT component composition approaches and architectures. Initial results show that current integration approaches mainly focus on performance evaluation of their integration solutions, which may be too narrow for fulfilling user goals by utilizing of IoT architectures.

The increasing digitisation impacts work and life. Process virtualization theory (PVT) offers a new, yet largely empirically unsupported theory to determine which processes are amenable to virtualisation. The paper at hand applies the theory to virtualisation that is established through the internet and commonly referred to as digitisation. As one of the most impacted areas for digitisation, we focus on the area of retail banking. A quantitative study among 368 German retail banking customers examines the PVT-based research model. The results indicate that sensory requirements and control requirements are the two strongest inhibitors for intended digital process use. The empirical examination of the rarely tested PVT contributes to a more thorough understanding how process-specific factors influence the digitisation in retail banking. Practitioners can use our results to increase the rate of digital transactions by addressing the suggested inhibitors of PVT. As a result, increased digitisation can improve the operational efficiency.

High-fidelity GUI prototyping provides a meaningful manner for illustrating the developers' understanding of the requirements formulated by the customer and can be used for productive discussions and clarification of requirements and expectations. However, high-fidelity prototypes are time-consuming and expensive to develop. Furthermore, the interpretation of requirements expressed in informal natural language is often error-prone due to ambiguities and misunderstandings. In this dissertation project, we will develop a methodology based on Natural Language Processing (NLP) for supporting GUI prototyping by automatically translating Natural Language Requirements (NLR) into a formal Domain-Specific Language (DSL) describing the GUI and its navigational schema. The generated DSL can be further translated into corresponding target platform prototypes and directly provided to the user for inspection. Most related systems stop after generating artifacts, however, we introduce an intelligent and automatic interaction mechanism that allows users to provide natural language feedback on generated prototypes in an iterative fashion, which accordingly will be translated into respective prototype changes.

Understanding the function learned by a neural network is crucial in many domains, e.g., to detect a model’s adaption to concept drift in online learning. Existing global surrogate model approaches generate explanations by maximizing the fidelity between the neural network and a surrogate model on a sample-basis, which can be very time-consuming. Therefore, these approaches are not applicable in scenarios where timely or frequent explanations are required. In this paper, we introduce a real-time approach for generating a symbolic representation of the function learned by a neural network. Our idea is to generate explanations via another neural network (called the Interpretation Network, or I-Net), which maps network parameters to a symbolic representation of the network function. We show that the training of an I-Net for a family of functions can be performed up-front and subsequent generation of an explanation only requires querying the I-Net once, which is computationally very efficient and does not require training data. We empirically evaluate our approach for the case of low-order polynomials as explanations, and show that it achieves competitive results for various data and function complexities. To the best of our knowledge, this is the first approach that attempts to learn mapping from neural networks to symbolic representations.

Using languages with formalized semantics for automating component integration is a well-established research area. As a consequence, independently developed software systems can interact without the need for manual integration effort in a "plug-and-play" manner. However, such dynamic adaptive system architectures are not widely used in Industrial IoT scenarios. Practitioners mostly rely on informal, domain-specific standards as formal interface specifications tend to become highly complex quickly. Nonetheless, this results in high manual integration efforts as integration knowledge cannot be reused. Thus, interface specification should be tailored towards its case-based requirements. Interface specifications should only be created and persisted evolutionary after specific integration tasks with knowledge management techniques. The resulting knowledge-driven architecture composition enables integration knowledge reusability and can ultimately automate component integration.

Prior work on leadership in information systems development (ISD) teams has assumed that all developers are treated equally by their team leader and ignored the possibility that differentiated leader-member exchange (LMX) may be an important instrument for team leaders to influence self-organizing, agile ISD teams. We conducted a concurrent mixed methods inquiry to understand how LMX differentiation is associated with developer satisfaction in agile ISD teams and through which team processes agile ISD teams address LMX differentiation. We ran a multilevel, multistage survey of 1,894 software developers in 217 teams and an embedded case study of five ISD teams drawing on qualitative data from 40 interviews of developers and team leaders. Two focus groups (one with 10 developers and one with 10 team leaders) helped to substantiate the meta-inferences from the quantitative and qualitative studies. The results showed that LMX differentiation was positively associated with developer satisfaction, especially in teams with high-quality team-member exchange (TMX). We identify three team processes (i.e., collectivization of resources, visible appreciation of privileges, and freeing up leader capacities) that are enacted through agile ISD practices and allow ISD teams to leverage benefits from LMX differentiation for all their members.

In this systematic literature review an overview of the recent developments in the field of Next Place Prediction is given. Next Place Prediction in this work refers to the prediction of where an individual human will go to next, based on continuous mobility data. It is therefore distinguished from other work in the field of next place prediction that uses, for example check-in data. This review aims to answer the following four questions: (1) Which features are used? (2) Which input data is required? (3) Which technique is used? (4) How is the prediction evaluated?

While the literature on consistent individual differences in correlated suites of physiological and behavioural traits is steadily growing for vertebrates, invertebrates have received less attention. The few studies that do exist have measured temporary physiological states (or responses), rather than consistent individual physiological traits. Here, I explore the consistency of individual differences in physiology and behaviour of n=53 shore crabs (Carcinus maenas) by repeatedly measuring haemolymph density (HD) and the crabs' responses to a novel environment. In crustaceans, HD is directly proportional to protein concentrations, and thus indicative of physiological condition. HD was highly repeatable, and crabs showed consistent individual differences in their behavioural responses to a novel environment, thus indicating individual consistency in both physiology and behaviour. Furthermore, HD was significantly correlated with the crabs' risk propensity, i.e. individuals with higher HD spent more time near shelter. Overall, this provides the first evidence for consistency in an endogenous physiological trait in an invertebrate. The link between consistent physiology and behaviour, i.e. coping styles, analogous to those found in vertebrates, suggests metabolic and/or immunological correlates of personality which offer great potential for future studies.

Purpose – In organizations, individual user’s compliance with business processes is important from a regulatory and efficiency point of view. The restriction of users’ choices by implementing a restrictive information system is a typical approach in many organizations. However, restrictions and mandated compliance may affect employees’ performance negatively. Especially when users need a certain degree of flexibility in completing their work activity. The purpose of this paper is to introduce the concept of directive explanations (DEs). DEs provide context-dependent feedback to users, but do not force users to comply. Design/methodology/approach – The experimental study used in this paper aims at investigating how DEs influence users’ process compliance. The authors used a laboratory experiment to test the proposed hypotheses. Every participant underwent four trials for which business process compliance was measured. Two trial blocks were used to cluster the four trials. Diagrammatic DEs were provided in one of the trial blocks, while textual DEs were provided in the other. Trial blocks were counterbalanced. Findings – The results of the experiment show that DEs influence a user’s compliance, but the effect varies for different types of DEs. The authors believe this study is significant as it empirically examines design characteristics of explanations from knowledge-based systems in the context of business processes. Research limitations/implications – This study is certainly not without limitations. The sample used for this study was drawn from undergraduate information systems management students. The sample is thus not representative of the general population of organizations’ IT users. However, a student sample adequately represents novice IT users, who are not very familiar with a business process. They are particularly suitable to study how users react to first-time contact with a DE. Practical implications – The findings of this study are important to designers and implementers of systems that guide users to follow business processes. As the authors have illustrated with a real-world scenario, an ERP system’s explanation can lack details on how a user can resolve a blocked activity. In situations in which users bypass restricted systems, DEs can guide them to comply with a business process. Particularly diagrammatic explanations, which depict actors, activities, and constraints for a business process, have been found to increase the probability that users’ behavior is business process compliant. Less time may be needed to resolve a situation, which can result in very efficient user-system cooperation. Originality/value – This study makes several important contributions to research on explanations, which are provided by knowledge-based systems. First, the authors conceptualized, designed, and investigated a novel type of explanations, namely, DEs. The results of this study show how dramatic the difference in process compliance performance is when exposed to certain types of DEs (in one group from 57 percent on the initial trial to 82 percent on the fourth trial). This insight is important to derive design guidelines for DE, particularly when multimedia material is used.

Trapezoidal Data Streams are an emerging topic, where not only the data volume increases, but also the data dimension, i.e. new features emerge. In this paper, we address the challenges that arise from this problem by providing a novel approach to restructure and prune Hoeffding trees. We evaluate our approach on synthetic datasets, where we can show that the approach significantly improves the performance compared to the baseline of an adjusted Hoeffding tree algorithm without restructuring and pruning.

BACKGROUND AND PURPOSE: Investigating the cost-effectiveness of future mobile stroke unit (MSU) services with respect to local idiosyncrasies is essential for enabling large-scale implementation of MSU services. The aim of this study was to assess the cost-effectiveness for varying urban German settings and modes of operation. METHODS: Costs of different operating times together with different personnel configurations were simulated. Different possible catchment zones, ischemic stroke incidence, circadian distribution, rates of alternative diagnoses, as well as missed cases were incorporated to model case coverage and patient numbers. Based on internationally reported clinical outcomes of MSUs, a 5-year Markov model was applied to analyze the cost-effectiveness for the different program setups. RESULTS: Compared with standard stroke care, MSUs achieved an additional 0.06 quality-adjusted life years (QALYs) over a 5-year time horizon. Assuming a catchment zone of 750,000 inhabitants and 8 h/7 day operation resulted in an incremental cost-effectiveness ratio (ICER) of €37,182 per QALY from a societal perspective and €45,104 per QALY from a healthcare perspective. Lower ICERs were possible when coverage was expanded to 16 h service on 7 days per week and larger populations. Sensitivity analyses revealed that missing ischemic strokes significantly deteriorated economic performance of MSU. CONCLUSIONS: Major determinants of cost-effectiveness should be addressed when setting up novel MSU programs. Catchment zones of more than 500,000-700,000 inhabitants and operating times of at least 12-16 h per day, 7 days per week could enable the most cost-effective MSU services in the German healthcare system.