Christoph G. Schuetz

Dual deep modeling : multi-level modeling with dual potencies and its formalization in F-Logic

An enterprise database contains a global, integrated, and consistent representation of a company’s data. Multi-level modeling facilitates the definition and maintenance of such an integrated conceptual data model in a dynamic environment of changing data requirements of diverse applications. Multi-level models transcend the traditional separation of class and object with clabjects as the central modeling primitive, which allows for a more flexible and natural representation of many real-world use cases. In deep instantiation, the number of instantiation levels of a clabject or property is indicated by a single potency. Dual deep modeling (DDM) differentiates between source potency and target potency of a property or association and supports the flexible instantiation and refinement of the property by statements connecting clabjects at different modeling levels. DDM comes with multiple generalization of clabjects, subsetting/specialization of properties, and multi-level cardinality constraints. Examples are presented using a UML-style notation for DDM together with UML class and object diagrams for the representation of two-level user views derived from the multi-level model. Syntax and semantics of DDM are formalized and implemented in F-Logic, supporting the modeler with integrity checks and rich query facilities.

Reference Modeling for Data Analysis: The BIRD Approach

Reference models for data analysis with data warehouses may consist of multidimensional reference models and analysis graphs. Multidimensional reference models are best-practice domain-specific data models for online analytical processing. Analysis graphs are reference models of analysis processes for event-driven data analysis. Small and medium-sized enterprises (SMEs) as well as large multinational companies may benefit from the use of reference models for data analysis. The availability of multidimensional reference models lowers the obstacles that inhibit SMEs from using business intelligence (BI) technology. Multinational companies may define multidimensional reference models for increased compliance among subsidiaries and departments. Furthermore, the definition of analysis graphs facilitates the handling of business events for both SMEs and large companies. Modelers may customize the chosen reference models, tailoring the models to the specific needs of the individual company or local subsidiary. Customizations may consist of additions, omissions, and modifications with respect to the reference model. In this paper, we propose a metamodel and customization approach for multidimensional reference models and analysis graphs. We specifically address the explicit modeling of key performance indicators as well as the definition of analysis situations and analysis graphs.

Multilevel Business Processes

ions

Ontology-based data description and discovery in a SWIM environment

ATM data in the upcoming SWIM originates from different authoritative sources (such as Eurocontrol) and is combined and enriched by different data providers (such as GroupEAD). It is further filtered, combined and enriched specifically by organizations (Airlines, ANSPs, Manufacturers) and provided to end-users for the accomplishment of specific tasks (such as the data in an EFB for a pilot conducting a specific flight). The relationships between these different data products are hidden in APIs or non-standardized interfaces of web services. This makes the discovery and combination of data products an intricate task requiring intervention of IT personnel. The lack of a common model to describe the contents of data products prevents automatic identification of relevant data products for a specific task.

Semantic data containers for realizing the full potential of system wide information management

In order to unleash the full potential of System Wide Information Management (SWIM), the BEST project (Achieving the Benefits of SWIM by Making Smart Use of Semantic Technologies) proposes the semantic container approach which shields service and application developers from the complexities of data provisioning in Air Traffic Management (ATM). In combination with SWIM, semantic containers facilitate the emergence of a marketplace of value-added information services, and allow for complex derivation chains of data sets. Along these derivation chains, existing data are intelligently filtered and prioritized as well as combined and annotated with additional information.

Modification Operations for Context-Aware Business Rule Management

The increasing number, complexity, and variability of business rules in todays enterprises introduces the need for their effective and flexible management. In several research fields, such as the semantic web, library science, and data tailoring, effective organization of knowledge is enabled by contexts. We previously proposed a static structure model for context-based business rule management. To enable flexibility, we complement this model by atomic and composed modification operations. Each atomic modification operation is associated with one of four roles: rule repository administrator, rule developer, user, and domain expert. This enables effective separation of tasks and responsibilities promoting efficient rule management. Composed modification operations describe combinations of atomic modification operations relevant in practice. We apply the proposed approach to the real-world use case of classifying aeronautical messages.

Semantic OLAP Patterns: Elements of Reusable Business Analytics

Online analytical processing (OLAP) allows domain experts to gain insights into a subject of analysis. Domain experts are often casual users who interact with OLAP systems using standardized reports covering most of the domain experts’ information needs. Analytical questions not answered by standardized reports must be posed as ad hoc queries. Casual users, however, are typically not familiar with OLAP data models and query languages, preferring to formulate questions in business terms. Experience from industrial research projects shows that ad hoc queries frequently follow certain patterns which can be leveraged to provide assistance to domain experts. For example, queries in many domains focus on the relationships between a set of interest and a set of comparison. This paper proposes a pattern definition framework which allows for a machine-readable representation of recurring, domain-independent patterns in OLAP. Semantic web technologies serve for the definition of OLAP patterns as well as the data models and business terms used to instantiate the patterns. Ad hoc query composition then amounts to selecting an appropriate pattern and instantiating that pattern by reference to semantic predicates that encode business terms. Pattern instances eventually translate into a target language, e.g., SQL.

Design, Management, and Customization of Data Analysis Reference Models Using Indyco Builder and Xquery

The design and implementation of analytical information systems requires both technical expertise and domain knowledge, the complexity of which often deters small and medium-sized enterprises from adopting business intelligence (BI) technology. Reference models for data analysis consist of domain-specific best-practice multidimensional models and greatly facilitate the development of analytical information systems. Definitions of business ratios, business terms, and frequent analysis situations are valuable additions to data analysis reference models which further lower the barriers that inhibit non-experts from using BI technology. Companies may then customize reference models to their specific needs. We demonstrate an approach which allows BI consultants, in collaboration with non-experts in BI technology, to employ off-the-shelf software for the design, management, and customization of data analysis reference models. We adopt Indyco Builder, a visual editor for dimensional fact models, for multidimensional reference modeling and provide a set of XQuery modules for the management and customization of the authored reference models.

Towards Scalability Guidelines for Semantic Data Container Management

Semantic container management is a promising approach to organize data. However, the scalability of this approach is challenging. By scalability in this paper, we mean the expressivity and size of the semantic data containers we can handle, given a suitable quality threshold. In this paper, we derive scalability characteristics of the semantic container approach in a structured way. We also describe actual experiments where we vary the number of available CPU cores and quality thresholds. We conclude this work-in-progress paper by describing how more measurements could be performed so that the missing guidelines could be provided.

Using superimposed multidimensional schemas and OLAP patterns for RDF data analysis

Abstract The foundations for traditional data analysis are Online Analytical Processing (OLAP) systems that operate on multidimensional (MD) data. The Resource Description Framework (RDF) serves as the foundation for the publication of a growing amount of semantic web data still largely untapped by companies for data analysis. Most RDF data sources, however, do not correspond to the MD modeling paradigm and, as a consequence, elude traditional OLAP. The complexity of RDF data in terms of structure, semantics, and query languages renders RDF data analysis challenging for a typical analyst not familiar with the underlying data model or the SPARQL query language. Hence, conducting RDF data analysis is not a straightforward task. We propose an approach for the definition of superimposed MD schemas over arbitrary RDF datasets and show how to represent the superimposed MD schemas using well-known semantic web technologies. On top of that, we introduce OLAP patterns for RDF data analysis, which are recurring, domain-independent elements of data analysis. Analysts may compose queries by instantiating a pattern using only the MD concepts and business terms. Upon pattern instantiation, the corresponding SPARQL query over the source data can be automatically generated, sparing analysts from technical details and fostering self-service capabilities.