Background

The reference points have to support the functions requested above. For example, micro-payment between two objects has to be reflected by specifying the information to be exchanged and the relationships to other objects, which might act as certifying instances, billing, or ac-counting server.

knowledge. In addition, Figure 22 shows the evolution of data, information, and knowledge.

Data forms the basis for information. If data is evaluated in a context, then data gets a meaning and becomes to information. Knowledge is extracted from interpreted information.

Figure 23 shows a distinction into different knowledge levels. The knowledge levels are similar to interpret as the evolution steps in Figure 22. A new level on top of knowledge is called meta-knowledge. Meta-Knowledge is knowledge about knowledge, for instance structural or strategic knowledge of a domain.

Meta-Knowledge Knowledge Information

Data

Figure 23: Knowledge Levels

The border between data, information, knowledge, and meta-knowledge is not fixed. The classi-fication depends on viewpoint and background. The following example [MS] shows the correla-tion between data, informacorrela-tion, and knowledge.

Example – digital pictures: While on the data level only bit streams are represented, the infor-mation level may contain additional format descriptions (especially those that identify the data as being a picture). Different information may be derived from the same data. On the knowledge level there may be semantic descriptors identifying the type of the picture (e.g. a landscape).

Searching for landscape pictures in a database would have no result. The information system may select pictures from the database and only on the knowledge level, a landscape painting could be distinguished from a portrait.

Knowledge Engineering

As it was described in the example above, data, information and knowledge are represented in different forms. Data is represented as a bit stream or character stream. Information is repre-sented as the data stream and its structure. Research in the area of Artificial Intelligence tried to find descriptions for knowledge. It was attempted to build systems, which contain and process knowledge, called Knowledge-based Systems (KBS) [W-KBS], e.g. expert systems. The first approaches attempted the knowledge of experts to be transferred into rules and data structure.

The beginning success from small academic prototypes was not transferable into a large com-mercial system. Knowledge Engineering shifted the paradigm from the ‘transfer approach’ to the ‘modeling approach’. Knowledge Engineering ‘turned the process of constructing KBSs from an art into engineering discipline’ [SBF, Hei]. The change of the paradigm has the follow-ing consequences [SBF]:

- like every model, such a model is only an approximation of the reality

- the modeling process is a cyclic process

- the modeling dependents on the subjective interpretations of the knowledge engineer Some outcomes of Knowledge Engineering are described in the next sections.

Principles

A set of design criteria for ontologies are described in the following. They are used to develop ontologies. This list was extracted from the articles [Per, Grub] by A.G. Perez.

- Clarity and Objectivity, the ontology should provide the meaning of defined terms by providing objective definitions and also natural language documentation

- Completeness, a definition expressed by a necessary and sufficient conditions is preferred over a partial definition

- Coherence, to permit inferences that are consistent with the definitions.

- Maximize monotonic extendibility, new general or specialized terms should be included in the ontology in such a way as does not require the revision of existing definitions.

- Minimal ontological commitments, making as few claims as possible about the world being modeled, which means that the ontology should specify as little as possible about the meaning of its terms, giving the parties committed to the ontology freedom to specialize and instantiate the ontology as required.

- Ontological Distinction Principle [Borgo], which means that classes in an ontology should be disjoint. The criterion used to isolate the core of properties considered invariant for an instance of a class is called the Identity Criterion.

- Modularity [Bernaras], minimize coupling between modules.

- Minimize the semantic distance between sibling concepts [Bernaras]. Similar concepts are grouped and represented as subclasses of one class and should be defined using the same primitives, whereas concepts, which are less similar, are represented further apart in the hi-erarchy.

- Standardize names whenever is possible

Types

Types differentiate in the amount and type of structure in the subject of the conceptualization.

The first dimension has three categories, which can be seen as different quality levels of ontolo-gies. Quality means the data, which is represented of it, e.g. data, information, and knowledge.

[Hei]

- Terminological ontologies such as lexicons specify the terms that are used to represent knowledge in the domain of discourse.

- Information ontologies, which specify the record, structure of databases. Conceptual schemata of databases are an example of this class of ontologies.

- Knowledge modeling ontologies specify conceptualizations of the knowledge. Compared to information ontologies knowledge modeling ontologies usually have a richer internal structure.

The second dimension has four categories. It differentiates the subjects of the conceptualization.

- Application ontologies contain all the definitions that are needed to model the knowledge required for a particular application. Typically, application ontologies are a mix of concepts that are taken from domain ontologies and from generic ontologies (which are described be-low). Moreover, application ontologies may contain method- and task specific extensions.

Application ontologies are not reusable themselves. They may be obtained by selecting theories from the ontology library, which are then fine tuned for the particular application.

- Domain ontologies express conceptualizations that are specific for particular domains.

Whereas the domain knowledge describes factual situations in a certain domain (e.g. chest pain is a manifestation of atherosclerosis), the domain ontology puts constraints on the structure and contents of domain knowledge (e.g. diseases have findings as manifestations).

- Generic ontologies are similar to domain ontologies, but the concepts that they define are considered generic across many fields. Typically, generic ontologies define concepts like state, event, process, action, component etc. The concepts in domain ontologies are often defined as specializations of concepts in generic ontologies.

- Representation ontologies explicate the conceptualizations that underlie knowledge repre-sentation formalisms. They are intended to be neutral with respect to world entities. That is, they provide a representational framework without making claims about the world. Domain ontologies and generic ontologies are described using the primitives provided by representa-tion ontologies.

The classification of an ontology, which is used to utilize knowledge for a system, can be help-ful to find a suitable description language respectively description format.

Components

Knowledge in ontologies is formalized using four kinds of components: classes, relations, func-tions, and instances.

- Classes are used in a broad sense. They can be abstract or concrete, elementary (electron) or composite (atom), real or fictitious. A concept can be anything, about something is said and, therefore, could be the description of a task, function, action, strategy, reasoning process, etc.

- Relations represent a type of interaction between concepts of the domain. They are for-mally defined as any subset of a product of n sets, that is: R : C1xC2x ... xCn. Examples of binary relations are subclass-of and connected to.

- Functions are a special case of relations in which the n-th element of the relationship is unique for the n-1 preceding elements. Formally, functions are defined as: F : C1xC2x ...

xCn-1 Cn. Examples of binary functions are Mother-of and square, and an example of a ternary function is price-of-a-used-car that calculates the price of a second-hand car de-pending on the car-model, manufacturing date and number of driven kilometers.

- Instances are used to represent elements.

Components explained here, are similar to them, which are used in the object-oriented design and in the entity relationship modeling. That is not surprising, because they are also used to reflect a part of the reality (existing knowledge).