Catálogo de publicaciones - libros

Requirement engineering plays a vital role in the development of the software. The quality of the software being developed depends on the non-functional requirements, which are still not derived effectively due to the conflicts between them. This paper presents an approach to identify the non-functional requirements for a given usecase description from the domain model such as Unified Modelling Language class diagram and goal based questionnaires. This approach makes use of the domain model to find out the behaviour of the system and possible constraints for actors in the system. The non-functional requirement taxonomy and the user preferences are used to analyse the conflicts, which is resolved based on trade-off analysis by prioritizing the preference. The prioritization depends on the dominating non-functional requirements from the inference engine.

Any good research needs a strong foundation. This chapter lays these foundations for the book. One characteristic of original research is the lack of a standardized vocabulary. Therefore, we will first explain the terminology of ontology and alignment. Then the concept of similarity and its meaning for ontologies will be described, on both an abstract and a concrete level with examples. Only with a thorough understanding of the main expressions, it is possible to follow the ideas of the succeeding chapters.

In this book, ontology alignment is supposed to suit specific practical scenarios and not only a high-level theory. This was one of the basic claims presented in the beginning. For this reason, the next sections will examine concrete use cases and derive requirements for ontology alignment from them. As they make clear in which direction the research needs to head, these requirements are another important part of the foundations. Based on these findings, new approaches for ontology alignment will then be possible.

Ontology alignment touches many related fields. Data integration but also ontology integration are not new topics. To ensure the reuse of existing ideas an extensive investigation on related work is necessary before an own innovative approach is developed. This chapter starts with theoretic considerations on alignment. It will then focus on other actual approaches for ontology alignment. We will reach out to related fields, in specific the integration efforts of the database community, which has been approaching this issue for many years. A good overview of related work is also given in de Bruijn et al. (2006).

After the foundations, the next step in research methodology is the actual creative and innovative one. This central chapter of the book will now develop a new approach for ontology alignment based on the previous findings. In fact, this will be achieved through several elements. First, we will shape a general underlying process for alignment. Specific methods for each process step are then going to lead to a concise basic approach (Ehrig and Sure, 2004a). The modular composition makes it easy to follow the new ideas in this work. We will show that already existing approaches also fit this process very well. Furthermore, in later chapters different methods in the process are going to be substituted to better meet the previously identified requirements. An additional important section for proving the value of this work is a thorough evaluation. After we will have introduced the fundamentals of ontology alignment evaluation, we will directly evaluate the novel approach and show its strengths and weaknesses.

We have originally identified specific core requirements for different use cases. Based on our alignment approach from the previous chapter we will now step-wise address these requirements. Specific extended methods are presented in the following sections. These include efficiency increasing methods, machine learning aid, intelligent user interaction, and adaptive strategies for underlying use cases. Each section will also include an evaluation. Finally, we present a strategy combining all these individual improvements. This sophisticated approach will be the final result of all these efforts. It shall serve to practically align real world ontologies, being open for future extensions. We have motivated that such an approach is a key step towards knowledge integration in information systems.

After the creation of the new approaches, we now need to deploy them. Therefore, we will first show how these are implemented. This chapter is split into three implementations of the alignment approach: a basic infrastructure for ontology alignment and mapping (FOAM), Ontology Mapping based on Axioms (OMA), and the OntoStudio plug-in OntoMap. All of them have specific strengths, though the most important implementation undoubtedly is FOAM. We will refer to it in the subsequent practical applications.

This chapter will show the application of the developed alignment techniques within the project SWAP. After a short general description of the project, its two case studies will be presented each with respective sections on ontology alignment (Haase et al., 2004a; Tempich et al., 2004). These show that the developed ideas actually do carry over to practice.

A second project relying on the developed alignment approaches is SEKT. Again the overall project will be presented first. The chapter is then going to continue with the description of the individual case studies and their need for ontology alignment.

Due to the many facets of knowledge integration one cannot expect the goal of ontology alignment to be completed within the next years. This part, in first place, will give directions on extending the presented methodology. In specific, it will generalize the ontology alignment approach to general structure alignment. Further, it will allow more complex alignments than the current one-to-one identity alignments.