Data quality and especially the assessment of data quality have been intensively discussed in research and practice alike. To support an economically oriented management of data quality and decision-making under uncertainty, it is essential to assess the data quality level by means of well-founded metrics. However, if not adequately defined, these metrics can lead to wrong decisions and economic losses. Therefore, based on a decision-oriented framework, we present a set of five requirements for data quality metrics. These requirements are relevant for a metric that aims to support an economically oriented management of data quality and decision-making under uncertainty. We further demonstrate the applicability and efficacy of these requirements by evaluating five data quality metrics for different data quality dimensions. Moreover, we discuss practical implications when applying the presented requirements.

No abstract

Abstract:We present a probability-based metric for semantic consistency using a set of uncertain rules. As opposed to existing metrics for semantic consistency, our metric allows to consider rules that are expected to be fulfilled with specific probabilities. The resulting metric values represent the probability that the assessed dataset is free of internal contradictions with regard to the uncertain rules and thus have a clear interpretation. The theoretical basis for determining the metric values are statistical tests and the concept of the p-value, allowing the interpretation of the metric value as a probability. We demonstrate the practical applicability and effectiveness of the metric in a real-world setting by analyzing a customer dataset of an insurance company. Here, the metric was applied to identify semantic consistency problems in the data and to support decision-making, for instance, when offering individual products to customers.

Efficient business processes play a major role in the success of companies. Business processes are captured and described by models that serve, for instance, as a starting point for implementing processes in a service-oriented way or for performance analysis. To support process modelers via methods and techniques (e.g., algorithms) in an automated manner, several research fields such as process mining and automated planning of process models have emerged. In particular, the aim of the latter research field is to enable the automated construction of process models using planning techniques. To this end, an automated construction of control flow patterns in process models is necessary. However, this task currently remains a widely unsolved issue for the central patterns parallel split and synchronization. We introduce novel concepts, which, in contrast to existing approaches, allow the construction of complex parallelizations (e.g., nested parallelizations and parallelizations with an arbitrary length of path segments) and are able to identify the set of feasible parallelizations. Moreover, we propose an algorithm facilitating the automated construction of parallel splits and synchronizations in process models. Our approach is evaluated according to key properties such as completeness, correctness and computational complexity. Furthermore, both the practical applicability within several real-world processes of different companies in various contexts as well as the practical utility of our approach are verified. The presented research expands the boundaries of automated planning of process models, adds more analytical rigor to automatic techniques in the context of business process management and contributes to control flow pattern theory.

The rapid development of e-commerce has led to a swiftly increasing number of competing providers in electronic markets, which maintain their own, individual data describing the offered items. Recommender systems are popular and powerful tools relying on this data to guide users to their individually best item choice. Literature suggests that data quality of item content data has substantial influence on recommendation quality. Thereby, the dimension completeness is expected to be particularly important. Herein resides a considerable chance to improve recommendation quality by increasing completeness via extending an item content data set with an additional data set of the same domain. This paper therefore proposes a procedure for such a systematic data extension and analyzes effects of the procedure regarding items, content and users based on real-world data sets from four leading web portals. The evaluation results suggest that the proposed procedure is indeed effective in enabling improved recommendation quality.