1. Representation of Ontology using Classified Interrelated Object Model Mihika Shah CSE-IT Department, Nirma UniversityAbstract- II. ONTOLOGY Ontology representation is the essential part of the ontology Sharing common understanding of the structure oflearning progress. This paper aims to provide a straight-forward information among people or software agents is the commonbut efficient way to represent ontology. An effective data model purpose to develop ontologies. People develop ontology totechnology – Classified Interrelated Object Model (CIOM) – is share common understanding under a specific context, tointroduced and utilized to represent ontology. The maincomponents of ontology development are elucidated and reuse the existing information, to explicitly expressdescribed, including ontology classes and their hierarchy, class preconditions of this context, and to store this knowledgeattributes, and inter-classes relationships. This paper provides a physically.general purpose methodology to facilitate the advanced ontologytechnologies. A. WHAT IS ONTOLOGY? In philosophy, ontology is a study subject that research on theKeywords- Classified interrelated object model, nature of existence, their properties, and their relations. It hasontology, ontology representation, semantic database model been applied to many other subjects. In the context of computer and information sciences, the commonly agreed I. INTRODUCTION definition of Ontology is proposed by Tom Gruber: anIn recent years the development of ontologies – explicit ontology defines a set of representational primitives withformal specifications of the terms in the domain and relations which to model a domain of knowledge or discourse. Theamong them – has been moving from the realm of Artificial- representational primitives are typically classes (or sets),Intelligence laboratories to the desktops of domain experts. attributes (or properties), and relationships (or relations amongOntology learning greatly helps ontology engineers to class members). The key role of ontologies with respect toconstruct ontologies. Ontology has been involved into a new database systems is to specify a data modelling representationgeneration – internet-based presentation – the Semantic Web. at a level of abstraction above specific database designsAs ontology spreading to Internet, ontology becomes one of (logical or physical), so that data can be exported, translated,the core technologies for knowledge exchange and inter- queried, and unified across independently developed systemssystem sharing purposes. and services. Ontologies are used to specify standard conceptual Ontology and data models, such as Enhanced Entity-vocabularies in which to exchange data among systems, Relationship Model, are the same in terms of representingprovide services for answering queries, publish reusable domain knowledge. But there are still some significantknowledge bases, and offer services to facilitate differences between them. First, Ontology represents a higherinteroperability across multiple, heterogeneous systems and level of abstraction than data models. Data models are focusdatabases. on current context that it elaborates to express while ontologyAn effective data model technology – Classified Interrelated can be shared and reused among various contexts. Second, asObject Model (CIOM) – is introduced and utilized to ontology stands at a higher abstract level it covers a largerrepresent ontology. Why is a database modelling method used scale of information than data models. Several data models forto represent ontology? Some of the advantages are: diverse domains might share the same ontology. Finally, the1 operation of ontology is usually isolated from its context • To eliminate the need to request a special designed while the data models are heavily related to the operation development tools to deploy ontology applications environments. To provide a straight-forward but efficient way to represent ontology B. ONTOLOGY v/s CONCEPTUAL MODEL • To take advantages of current well-developed In the SE and IS communities, perhaps due to the historical database technologies importance of conceptual modelling, there is frequent confusion between ontology and conceptual models. In some • To use existing database to store complicated and sense, an ontology has a similar function to a database schema large-scale ontology information base. because the first provides meta-information that describes the semantics of the terms or data, but there are several important • To offer a general purpose methodology to facilitate differences between these concepts : the advanced ontology technologies. • Languages for defining and representing ontologies (OWL, In the beginning of this paper, the major traits of ontology etc.) are syntactically and semantically richer than commonis described and discussed. Then the most impressing features approaches for databases (SQL, etc.).of CIOM is briefed with a schema example. Finally, CIOM iscatered to represent the core building blocks of ontology. • The knowledge that is described by an ontology consists ofAdditionally, some CIOM schemas of ontology are provided semi structured information (that is, texts in natural language)for the purpose of better illustration. as opposed to the very structured data of the database (tables, classes of objects, etc).
2. • An ontology must be a shared and consensual Ontology is a methodology to formula the definition ofconceptualization because it is used for information sharing representational vocabulary for common sharing purpose.and exchange. Identifiers in a database schema are used Ontology is a specification of conceptualization. It calls for aspecifically for a concrete system and do not have the need to specific description of all kinds of entities, their properties andmake an effort to reach the equivalent of ontological their relations. There are several kinds of ontology languagesagreements. already developed to encode the ontology, such as Resource Description Framework and Web Ontology Language. These• An ontology provides a domain theory and not the structure technologies are special designed for ontology representation.of a data container. They might require familiarities of these technologies, exclusive development tools, and specially trained experts forC. ONTOLOGY PROCESS ontology development tasks. Furthermore, ontologies developed on these tools are only able to share within theHere is the step by step progress to create ontology with a same platform, which limits their usability. In this paper, agiven domain: more general and universal data model method CIOM is utilized to represent ontology. 1. Define the context within the given domain. 2. Create classes and their hierarchy. 3. Identify the attributes of classes. A. CLASSES AND THEIR HIERARCHY 4. Connect classes with inter-relationships among them. In the ontology, classes are defined to classify all kinds of existences. A class usually refers to a collection or a category of objects sharing some common character and well accepted III. CLASSIFIED INTERRELATED OBJECT MODEL under commonsense. All the objects under this category are normally named as “instances” of this class. Objects of theSemantic Database Model (SDM) is a high-level semantics- same class are also differentiated themselves by their ownbased database description and structuring formalism for traits. This diversity implies that the objects are organized indatabase systems. SDM captures more meaning of the real hierarchy. The objects sharing the same trait of a class areworld. Comparing to other data model technologies, SDM grouped as instances of a subclass of this class. When classesoutstands itself by its expressivity and effectiveness. and their hierarchy are represented in CIOM, an ontologyClassified Interrelated Object Model (CIOM) is a simplified class is an oval with a class name written inside, while asubset of SDM, with basic structures, operations, and subclass is also a class but drawn as a double-line arrowconstraints of SDM. pointing from its parent class. An ontology class “Vehicle” The CIOM we use throughout this paper is loosely based represented with CIOM is shown in Figure 2. A “Vehicle” hason a SDM data model presented in “Database Description two types of subclasses: “Car” and “Truck”, while A “Car”with SDM: A Semantic Database Model”. CIOM primarily has two types of subclasses: “Sedan” and “Sports Utilityconsists of classes, subclasses, and member attributes. A class Vehicle (SUV)”.is defined as a collection of entities, with a class name toidentify itself from others. In CIOM, an oval with a classname written inside is denoted as this class. A subclass meansspecialization, viz. its membership is a subset of the membersof its parent class. Member attributes are the common aspectsof members of a class. In CIOM, an attribute is drawn as apair of arrows pointing from one class to another withopposite directions. A simple CIOM schema for a class“Vehicle” is shown in Figure 1. As shown in this figure, a“Vehicle” has a unique “Vehicle Identification Number(VIN)”. The type of VIN is “String”, which implies that theVIN consists of alphanumeric characters. A “Vehicle” has twotypes of subclasses: “Car” and “Truck”. Figure 2- An ontology class ‘Vehicle’ represented using CIOM. B. CLASS ATTRIBUTES For an ontology class, its attributes are described as all the related ontology classes, typically some built-in ontology classes. These attributes are those sharing traits that identify the class itself from other classes. The representing built-in classes include String, Number, Date, and other atomic classes. String is the collection of all alphanumeric characters. Number isFigure 1- A CIOM schema for a class ‘Vehicle’ the collection of all numeric values of digits. Date is the collection of all time entities in a calendar system. Cardinality is a measure of the number of the IV. REPRESENT ONTOLOGY corresponding attributes an ontology class has. The cardinality of an ontology attribute is quite different to
3. a data model attribute. Normally speaking, the cardinality of a data model attribute is categorized into three types: One-to-One, One-to-Many, and Many-to- Many. Their meanings of these types are self- explanatory. However, an ontology attribute does not has Many-to-Many cardinality as ontology stresses a formulized conceptualization the while data model technologies emphasize on the representation of all entities. As a result, the inverse attributes of an ontology class is always has a monotony cardinality. For example, a person has attributes, such as “Social Security Number (SSN)”, “Name”, “Date of Birth Figure 4 - The inter-classes relations of a vehicle represented (DOB)”, and “Phone Number”. A person can have with CIOM several phone numbers at the same time, while a phone number can belong to several persons. For an ontology V. CONCLUSION class “Person”, it is only focused on the main target In this paper, a general purpose method, CIOM, instead of “Person” while arguing its several objects of this class specific ontology language and develop tools, is used to is meaningless as ontology is an abstraction at represent ontology. The definition and major components of conceptual level. The attributes of an ontology class ontology are briefly discussed in the beginning. Then the most “Person” represented with CIOM is shown in Figure 3. outstanding features of CIOM are also brought into When drawing a CIOM ontology class, just simply instruction. Finally, this paper illustrates the progress to use neglect the inverse attributes or assign their cardinality CIOM to represent the major component of ontology. values to one. The potential applications of this method include representing ontology with a general purpose modelling technology, such as EER, UML, and CIOM; storing ontology on general database, such as MySQL, DB2, and Oracle; and facilitating the sharing of ontology via general purpose tools, such as Eclipse, PowerBuilder, and Visual Studio. Despite of the express power of CIOM, only part of CIOM features are applied in this paper. CIOM also provides operation level modelling. CIOM supports: using predicates, operations, and set-operators to define subclasses; class attribute mapping; class grouping. VI. REFERENCES 1. Emdad Ahmed. Use of Ontologies in SoftwareFigure 3- The attributes of a class ‘Person’ represented using Engineering (Page 1-4)CIOM 2. Dragan Gasevic, Nima Kaviani, Milan Milanovi. Ontologies and software engineering. (Page 1-9)C. INTER-CLASSES RELATIONS 3. Eva Blomqvist, Kurt Sandkuhl. Patterns in ontology A relation between two ontology classes interprets how the engineering : classification of ontology patterns. two classes, more precisely the objects of these classes, are 4. Yair Wand, Carson Woo. Ontology based rules for related. Typically a relation is a particular connection object oriented enterprise modelling. between two classes specifies how an object is connected to the other in ontology. The CIOM gains its expression power by providing an effective way for relation description. CIOM ascribes a relation between two classes as a special kind of attribute, denoted as Class Attribute. The only visible difference might be that attributes are built-in classes while the classes within relation category are abstract defined classes, normally customized by users. However, it is this difference that renders the efficiency of semantic expression of ontology. The cardinality of a relation is categorized into three types: One-to-One, One-to-Many, and Many-to-Many. Unlike class attributes, all of them are fully supported within the same relation category. For example, a vehicle is equipped with only one engine while an engine can only equip one vehicle; a vehicle has only one owner while a person can have several vehicles at the same time; a vehicle can only be manufactured by one manufacturer while a manufacturer can produce server types of vehicles.