Integrating Ontology Models and Conceptual Models using a Meta Modeling Approach

The integration of ontologies in enterprise information systems has recently gained a lot of interest. In this paper we show three approaches to integrate models of ontologies and conceptual models of information systems. Therefore we revert to a meta modeling approach for specifying the modeling languages. The approaches have been implemented using the ADONIS meta modeling platform and the Protégé ontology editor and knowledge acquisition system. Introduction The use of web-based ontologies and their contribution to business innovation has received a lot of attention in the last years, e.g. (Cardoso, Hepp, and Lytras 2007). Examples can be found in different areas of enterprise information systems such as enterprise content management, enterprise information integration or enterprise service bus (Kobielus 2007). To analyze the relationships between business and technology the use of conceptual models has been shown to be a successful approach (Mylopoulos 1992). Thereby, the complexity of these relationships can be reduced and human understanding can be improved. Several approaches have been discussed to enhance both the act of creating conceptual models as well as the execution of the models by using semantic schemata, e.g. in the area of business process management (Lautenbacher, Bauer, and Seitz 2008; Hoefferer 2007). This allows for example to implement semantic search functionalities for process models, automate process executions by using semantic web service descriptions or realize auto-completion functionalities during the design phase (Lautenbacher, Bauer, and Seitz 2008). Although the conceptual foundation of these approaches shows several similarities, the actual realization on various technical platforms is sometimes not discussed in detail. To realize the alignment of conceptual models and semantic schemata on a technical level, it is however essential to elaborate on the details of these relationships. This allows to assess the concrete benefits or drawbacks of a particular approach. In the paper at hand we focus on the integration of ontology models and conceptual models by using a meta modeling approach. We will describe three possible approaches for a technical integration and discuss pros and cons of each approach. Conceptual Modeling In the area of information systems (IS) conceptual models are today widely used to facilitate the development, implementation, use, and maintenance of IS (Wand and Weber 2002). Conceptual modeling can be described as the ”activity of formally describing” aspects of the physical and social world for the purposes of human understanding and communication (Mylopoulos 1992). During the phases of development and implementation, conceptual models can be used to support the process of requirements engineering. Furthermore, conceptual models facilitate tasks such as the elicitation, negotiation, documentation, and validation of requirements. This permits to detect and correct possible errors at an early stage even prior to implementation (Pohl 1997; Wand and Weber 2002). Several proposals have been made for conceptual modeling techniques and languages – for an overview and discussion see (Roland and Prakash 2000). One of the advantages that are seen in using conceptual modeling is that it captures the semantics of an application through the use of a formal notation (Mylopoulos 1992). However, the descriptions resulting from conceptual modeling are intended to be used by humans and not machines. The semantics contained in conceptual models are therefore to a large part implicit and cannot be processed. With the upcoming of web-based semantic schemata such as the Web Ontology Language (OWL) two opportunities have emerged for the extension of conceptual models: On the one hand the semantic schemata need to be considered when designing and implementing information systems. This aspect can be represented in conceptual modeling and thus allows for an analysis of these semantic technologies in the context of information systems (Fill 2009a). On the other hand, using shared semantic schemata during the activity of conceptual modeling itself can improve the creation and use of conceptual models (Fonseca and Martin 2007). In particular, additional parts of the implicit semantics being contained in the models can be made explicit and used for processing. Meta Modeling Approach To describe the conceptual models in a formal notation we build upon a meta modeling approach for specifying the modeling language (Karagiannis and Kuehn 2002). Karagiannis and Kuehn divide modeling methods into a modeling technique and mechanisms and algorithms (see figure 1). The modeling technique comprises a modeling language and a modeling procedure. The modeling language is used to describe the models and is itself split into syntax, semantics, and a graphical notation. The semantics of the modeling language refer to a semantic schema and contain a mapping of the syntax of the modeling language to the schema. By separating the graphical notation from syntax a greater flexibility for the adaptation of the graphical representation is achieved (Fill 2009b). The modeling procedure defines the way how to apply the modeling language. Mechanisms and algorithms are used by the modeling procedure. They may either be generic, i.e. applicable to all modeling languages, specific, i.e. applicable only to particular modeling languages or hybrid, i.e. with certain parts being generic and certain parts being specific or adaptable. Based on these relationships a meta model can now be viewed as a model of a modeling language (Favre 2005).