| |
UKRAINIAN | |
RUSSIAN |
In present time the task of forming the conceptual "transparent" representations for poorly – structured domains is a very urgent one.
Today the leading paradigm of structuring information flows are ontologies or hierarchical conceptual structures, which are formed by an analyst on the basis of study and structuring the flow of information, documents, protocols, knowledge extraction, and other sources.
Ontology Engineering (OE) develops the fundamental knowledge of engineering – the science of the models and methods of knowledge extracting, structuring and formalization.
Actually knowledge engineering is a branch of artificial intelligence, while the OE covers a wider range of applications – from knowledge management systems to distance learning.
Ontological engineering is taking its first steps; therefore each analyst goes through trials and errors, creating complex ontological structures, reflecting the maze of professional knowledge in various application areas.
The ground for master's work topic is the relevance of constructing knowledge bases on the basis of ontology, which allow working with multi-level models of space and time.
Such classification of properties is necessary for regular studying of properties ontology and their applications in practice ontology which covers properties of the most different ontology, without limiting researchers in detection of new properties.
Therefore the work purpose is working out of classification of known properties from the literature ontology, convenient for regular estimation ontology in practice. Such classification is necessary for a choice most a totality of properties at estimation any ontology depending on "type" ontology and from the pursued purposes.
Such classification can be taken as a principle the uniform approach to estimation and definitions in unequivocal terms of properties ontology, and then and to objective estimation ontology. The received classification should be used as a source of requirements to use of the ontologic approach in practice in the field of SAPR.
Today the artificial intellect is an extensive area of researches and workings out of the intellectual systems intended for work in difficultly formalizable spheres of activity of the person. Now it is accepted to allocate some directions of development of an artificial intellect: one of directions is connected with working out of the intellectual systems based on knowledge. In the given direction are engaged in working out of models of representation of knowledge, creation of knowledge bases.
One of perspective applications of methods of the given direction is creation of intellectual training systems (ITS). ITS are intended for automation and a training individualization. Traditionally allocate four models of representation of the knowledge used at construction of knowledge bases of systems based on knowledge:
1. Logic models;
2. Produktsionnye models;
3. Frame models;
4. Semantic networks.
However at present a particular interest in researchers of an artificial intellect cause ontology. Ontology can be used for representation of knowledge of concepts of a subject domain and presumable relations between them, for the description of the maintenance of web pages. Besides ontology it is possible to use at construction of knowledge bases not only IOS, but also any other appendices.
Work assumes clearing of a situation with definition of concepts "ontology" and "property ontology", the analysis of existing classifications of properties and discussion of possibility of their use for regular estimation ontology and their models in practice. To present the classification of properties aimed at coverage and streamlining of all known properties from the literature, and also on "accumulation" earlier not discussed, but inherent ontology properties. Taking into account presence in structure ontology various kinds of communications to detail classification of various structural properties.
It is necessary to execute:
Classification ontology, as a source of the general requirements to effective construction ontology in various subject domains;
To list specific requirements to ontology in SAPR;
To offer the general approaches to construction ontology in SAPR.
1. Working out of the problem-focused graphic editor on system engineering ontology that answers multilevel model of space and time.
2. System engineering of input semantic production over set ontology.
Ontologies – one of the modern trends in the field of artificial intelligence. In general terms, ontology is defined as the knowledge base of a special form, or as a specification of conceptualization of the domain.
This means that in the domain based on the classification of basic terms the major concepts (concepts) are distinguished and connections between them are set out. This process is called conceptualization.
Then, ontology can be represented graphically or described in a formal language (formal ontology) – it is a process of ontology specification.
Ontologies can be used:
to describe the domains of scientific research,
to describe the academic disciplines and curricula,
to prepare library catalogs.
Introduction of library catalogs as a formal ontology will allow to automate their processing and to perform semantic search in the ontological space that describes a set of libraries.
There are different types of classification of ontologies. From my point of view, it would be most useful to distinguish two types of classification of ontologies:
Semantic:
1) according to the level of expressiveness;
2) according to the degree of formality;
3) according to the level of detail representations.
Pragmatic:
1) according to the degree of dependence on a specific task or application area;
2) by the language of ontological knowledge representation;
3) by domain;
4) for the purpose of establishment;
5) according to the content.
Here is a brief description of each classification.
Semantic classification
1. In terms of expressiveness
Heavy ontologies
Lightweight ontologies.
2. On the degree of formality
Informal;
More formal:
- Based on the terms;
- Based on the concept.
Strongly formalized.
3. The level of detail representations
Low;
High.
Pragmatic Classification
1. According to the dependence on a specific task or application domain
Upper level;
Oriented to the domain;
Task-oriented;
Applied ontologies.
2. In the language of ontological knowledge representations
RDF
DAML+OIL
OWL (Web Ontology Language)
KIF (Knowledge Interchange Format, or format of knowledge interchange)
CycL (a language for describing ontology Cyc)
OCML (Operational Conceptual Modeling Language)
LOOM и Power Loom
Ontolingua
F-Logic
3. By the domain
Ontology reflects the general knowledge of the domain, such as concepts class hierarchy and semantic relations in these classes. For each domain ontology is made by the experts in their field, who conduct the formalization of knowledge, definitions and rules of obtaining new knowledge.
4. For the purpose of establishment
Annex ontologies;
Reference ontologies.
There are different interpretations of the "ontology" concept. In this work, ontology is understood as a structural specification of a certain domain, its conceptual description in the form of formalized representation, which includes a glossary of terms in the domain and the logical expressions that describe the relationships of these concepts
On a wave of interest to ontologies there were created tools and mechanisms that specifically targeted to the wide use of ontologies in the problems of intelligent search, classifications and revelation of inconsistencies in the data, modeling the behavior of intelligent agents. However, even having a good tool environment does not eliminate the problems associated with the difficulty of designing and constructing ontologies themselves and automated ontology extraction process, as a whole task of knowledge extraction, is currently do not have its effective solution. The more valuable are already developed ontologies and experience of their usage for solving a wide range of tasks
In the process of creating modern intelligent information systems it is often required the integration of knowledge from disparate sources and, consequently, effective solution to problems associated with the replication of knowledge. Still there is no satisfactory solution to the problem of automating the process of selecting an adequate specificity of a particular domain and taken to it style of reasoning by the experts means of knowledge representation. That is why to this day researches are relevant developing the approach to the representation and reproduction of knowledge, which on the one hand would allow taking into account more adequately the domain specifics, and on the other – to represent and use knowledge in some unified manner.
Ontological models for the time of studies in this area have undergone considerable development. At the present time for creating and maintaining ontologies there are number of tools that in addition to general viewing and editing functions do support the documentation of ontologies, import and export ontologies in various formats and languages, support for graphic editing, management of ontologies libraries, etc.
These tools for ontologies building have several significant drawbacks. Most of the tools maintain its ontology in a text file, which limits the size of the ontology, have low productivity, needs further development of algorithms for the convenience of working with stored metadata, have redundant functions, which complicates the user experience.
1. Core Software Ontology. Core Ontology of Software Components. Core Ontology of Services [Электронный ресурс]. OntoWare Group. – Режим доступа: http://cos.ontoware.org
2. Gruninger Michael. An Ontology Framework / Michael Gruninger, Leo Obrst // Ontology Summit – NIST, (Gaithersburg, MD April 22 – 23, 2007). – Gaithersburg, 2007.
3. Гладун А.Я. Онтологии в корпоративных системах. Часть I / А.Я. Гладун, Ю.В. Рогушина // Корпоративные системы. – 2006. – № 1. – С. 41–47.
4. Глибовець М.М. Побудова україномовної онтології засобами СУБД / М.М. Глибовець, О.О. Марченко, А.О. Никоненко // Наукові записки. Комп’ютерні науки / Національний університет «Києво–Могилянська академія». - 2008. - Том 86. - С. 46-49.
5. Ontolingua [Электронный ресурс]. Artificial Intelligence Laboratory of the Department of Computer Science at Stanford University. – Режим доступа: http://www.ksl.stanford.edu/software/ontolingua/
6. Овдей О.М. Обзор инструментов инженерии онтологий / О.М. Овдей, Г.Ю. Проскудина // Электронные библиотеки. – 2004. – Т. 7, вып. 4. – С. 3–19.
7. Константинова Н.С. Онтологии как системы хранения знаний [Электронный ресурс] / Н.С. Константинова, О.А. Митрофанова. – Режим доступа: http://www.sci-innov.ru/icatalog_new/index.php?action=send_att&entry_id =68352&fname=68352e2-st08_(Митрофанова О.А.).pdf
8. Григорьев А.В. Семантика модели предметной области для интеллектуальных САПР. Научные труды Донецкого государственного университета. Серия «Информатика, кибернетика и вычислительная техника», (ИКВТ–2000) выпуск 10. – Донецк, ДонГТУ, 2000. – С. 148–154.
9. Григорьев А.В. Комплекс моделей САПР как система взаимосвязанных уровней о действительности. Научные труды Донецкого государственного университета. Серия «Информатика, кибернетика и вычислительная техника», (ИКВТ-2000) выпуск 10. – Донецк, ДонГТУ, 2000; С. 155–167.
10. Григорьев А.В. Вербальная модель предметной области для интеллектуальных САПР. Науковi працi Донецького Державного технiчного унiверситету. Серiя «Обчислювальна технiка та автоматизацiя». Випуск 20. – Донецьк, ДонДТУ, 2000. – С. 171–180.