관계형 데이터베이스로부터 온톨로지 생성을 위한 변환 규칙에 관한 연구

Xu Ni 전북대학교 2007 국내석사

Until now our insight into the documents and capabilities available are based on keyword searches stored in the form of database, abetted by clever use of document connectivity and usage patterns in the World Wide Web. It does not provide any means of the talking about the semantics of data. For the sake of the Semantic Web to structure the document with semantic way, it is important to translate the document stored in the form of database currently to Ontology. So it is important and necessary to map database to ontology to implement upgrading the World Wide Web composed of the static web pages to the semantic web structured by ontology. In recent years, there are three approaches have been reported. The first one is based in the semi-automatic generation of ontology from Relational DB model [4, 9]. Then mappings are defined between the DB and the generated ontology. The second approach, proposes the manual annotation of dynamic Web pages which publish DB content, with information about the underlying DB and how each content item in a page is extracted from the DB [4, 10]. The third approach tries to map an existing DB to an appropriate existing ontology implemented in RDF(S) or OWL with the declarative language R2O[4]. But the first one does not allow the population of an existing ontology, which is a big limitation, the second approach does not deal neither with complex mapping situations and assumes we want to make our database schema public, which is not always the case although the third one R2O is automatically generating language, it does not solve the definition of degrees of similarity between DB elements and ontology very well and the readability of it can not be controlled easily by the users. Generally it is difficult for most people to build the ontology based on the database knowledge base very well. Therefore, there needs the method for converting or mapping between RDB and ontology for providing any means of the talking about the semantics of data. This paper proposes the rules to translate from Database to Ontology in the domain specified Relational Database (RDB). The approach according to the characteristics of the elements of database such as table, record, attribute, primary key, foreign key, and relationship, generates the elements of ontology in OWL: Class, SubClass, Objectproperty, Datatype property, individual, and cardinality. 현재 인터넷은 시멘틱 웹(Semantic Web)으로의 진입단계에 있다. 기존의 연구들은 데이터베이스에 저장된 정보를 기반으로 관계를 도출하여 사용자들의 검색 요구에 충족하는 결과를 제공하기 위해 노력해왔다. 그러나 이는 데이터에 대한 의미적 관계를 통한 서비스를 제공하지 못하는 문제를 안고 있다. 이와 같은 문제의 해결을 위해 과거에도 시소러스(Thesaurus), 워드넷(Wordnet)과 같은 도구들을 정보 구축 단계에서 어휘선정에 적용하였으나 검색 효율성 향상에는 미흡하였다. 이는 인간의 사고체계에서 용어간의 관계 뿐 아니라 의미 인지를 통한 추론의 기능까지를 수용하지 못하기 때문이다. 이러한 문제는 온톨로지(Ontology)를 통하여 해소될 것으로 기대되고 있다. 따라서 의미적 관계를 표현 가능한 시멘틱 웹을 위해서는 데이터베이스에 저장된 정보를 분석을 톨하여 온톨로지로 변환하는 것은 중요한 일이다. 최근 들어, 데이터베이스를 온톨로지로 변환하기 위한 많은 연구들이 진행되고 있으며, 접근 방식은 크게 세 가지 정도로 분류되고 있다. 첫째는, 관계형 데이터베이스로부터 반자동으로 온톨로지도 변환하는 것이고[4,9]. 둘째는, 메타정보의 기술을 통한 수동적인 방법이다[4, 10]. 셋째는 R2O 언어를 이용하여 RDF(Resource Description Framework)또는 OWL(Web Ontology Language)로 구현된 기존 온톨로지에 매핑하는 방법이다[4]. 그러나 이러한 방법들은 수작업을 피할수 없으며, 일반적으로 사람이 직접 지식베이스를 구축하는 일은 매우 어려운 일이다. 본 논문은 관계형 데이터베이스로부터 온톨로지도 변환하기 위한 규칙을 제안한다. 이를 위해, 데이터베이스의 요소인 테이블(Table), 레코드(Record), 어트리뷰트(Attribute), 프라이머리키(Primary Key)등의 특성을 분석하고, 또한 OWL의 요소들인 클래스(Class), 서브클래스(SubClass), 오브젝트 프로퍼티(Objectproperty)등의 특성을 분석하였다. 이를 통하여 RDB를 OWL로 변환하기 위한 기본적인 규칙들을 정의하였다.

ONTOLOGY BASED GEOSPATIAL MODEL FOR PERSONALIZED ROUTE FINDING

아볼가셈 인하대학교 대학원 2008 국내박사

This research addresses an ontology-based architecture to design a user-centric ontology-based cost model for route finding analysis using a Multi-Criteria Decision Making (MCDM) technique in a Geographic Information System (GIS). GIS is computer-based system, which allows storage, editing, maintenance, presentation, and access to spatial data. A cost model is an essential element of the route finding analysis in GIS. User-centric (personalized) model provides an outcome based on the user’s preference. There is some dissatisfaction regarding existing route finding architectures and their cost models which are based on a one-dimensional and fixed criterion such as distance. With this in mind, the objectives of the present study are to develop an ontology-based architecture and a user-centric cost model for route finding analysis using MCDM-Analytic Network Process (ANP) technique. This architecture is completely flexible, dynamic, reusable, sharable and interoperable. The cost model employs several additional user-centric quantitative and qualitative criteria using an ontology concept. In order to address the research contributions, after a short review of the significance of a user-centric view, the user-centric ontology-based route finding architecture is proposed. This architecture includes four components, the user-centric ontology-based cost model, the ontology-based route finding analysis, geospatial component, and the data repository component. The cost model involves the cost model ontology and cost model task ontology. To create the cost model ontology, four stages are introduced including cost model domain ontology, conceptualization, formalization, and implementation. The appropriate criteria were determined after designing the cost model ontology. For weighting and integrating these criteria, the cost model task ontology using MCDM-ANP technique was designed. Finally, the ontology-based route finding analysis was developed including route finding ontology, route finding task ontology. The procedure for route finding ontology creating methodology is similar to cost model ontology including four mentioned steps. The route finding task ontology were developed employing the route finding algorithm and the route finding inference engine. Therefore, the result of this architecture is based on the user's preferences and the contextual situation around user’s position which will improve user satisfaction. The user-centric ontology-based cost models for different scenarios in three types of trips (tourist, particular, and business trip) were determined. The derived user-centric ontology-based cost models of this research are the reusable and sharable capability of the ontology-based route finding analysis. This makes it possible to accept any road database with any format and structure connected to these ontology-based architectures in order to perform the route finding processing. The user-centric ontology-based models as well as all POIs information in this work were successfully implemented in a Seoul road network. The user-centric results and the model evaluations on the real data illustrate the merits of the present approach. All of the paths determined through the route finding analysis in the study area corresponded well with the paths that a traveller would normally choose in reality. Nevertheless, there are few limitations for this research, for example, the computational load using an OWL (Ontology Web Language)-based database for route finding analysis is higher than employing a typical database (e.g. Oracle), as a consequence of the large size of the OWL file. However, with the advent of increasingly powerful data compression and data storage techniques, this problem can be remedied.

전자상거래 통합을 위한 온톨로지 구축

장태우 서울대학교 대학원 2004 국내박사

기업간 전자상거래에 관련한 정보기술의 생명주기가 짧아지고 인력 및 프로젝트의 특성이 다양화되어 가는 등 정보통신 시장환경이 변화하고 있다. 또한 이질적이고 분산되어 있는 전자상거래 상의 정보들은 그 내용 또한 기계가 이해하기에는 어려움이 많다. 이러한 시장환경과 정보기술 상의 어려움을 극복하기 위해서는 비용요소를 줄일 수 있는 자기기술적(self-describing) 규약이나 소프트웨어 설계 방법들이 필요하며, 그 중 의미론적인 통합을 위한 온톨로지(Ontology)의 구축에 대한 필요성이 제기되고 있다. 본 연구에서는 정보시스템 통합 및 온톨로지에 대한 기존의 연구들에 대해 정리하고, 전자상거래 통합 시스템을 구축하기 위해 필요한 협업 프레임워크와 이를 가능하게 하는 구문론적·의미론적 통합 및 온톨로지와 에이전트를 이용한 문제 해결 방법을 제시한다. 또한 문제 해결의 핵심이 되는 전자상거래를 위한 온톨로지를 구축하기 위해 10단계로 구성되는 시스템공학적 방법론을 제안하고, 방법론에 따라 RDF(Resource Description Framework)와 RDF Schema(RDFS), DAML(DARPA Agent Markup Language), PSL(Process Specification Language)을 이용하는 새로운 해결책을 모색하고자 한다. 온톨로지를 구축하기 위한 메타모형은 ISO/IEC(International Standardization Organization and International Electrotechnical Commission)에서 제정한 IRDS(Information Resource Dictionary Standard, ISO/IEC 10027:1990) 프레임워크를 기본으로 하여 계층적으로 구성하였다. 자원에 대한 데이터 모형화와 PSL-ontology의 분류 및 체계화는 RDF/RDFS 및 DAML을 사용하여 세 계층으로 구성하고, 프로세스 모형화를 위해 전자상거래에서 사용되는 프로세스 용어를 DAML을 이용하여 PSL-ontology의 하위 계층으로 종속시켜 네 계층으로 체계화하였다. 관련 자원을 포함한 전체 프로세스 구성은 XML(extensible Markup Language)로 표현한 PSL을 이용함으로써 분산된 전자상거래 환경의 데이터와 프로세스를 통합적으로 표현할 수 있게 된다. 데이터베이스 또는 데이터 사전 등의 기본적인 의미론 체계가 이미 구축된 시스템에 대해서는 이종 데이터베이스에 대한 통합 스키마와 기초적 온톨로지 구축에 대한 방법을 제시하고 온톨로지 간의 통합에 대한 연구의 기반을 마련하도록 한다. 실례로 견적 및 구매주문과 관련한 전자상거래 시나리오를 제시하고, 로제타넷(RosettaNet)의 표준을 참고한 프로세스 및 데이터에 대해 제시한 방법론에 따라 구축한 온톨로지를 적용하여 통합가능성을 제시한다. 마지막으로, 온톨로지를 사용하고 평가하기 위한 방법론으로써의 검색과 추론, 평가 기능에 대해 개략적으로 설명한 후, 데이터베이스를 활용한 프로토타입을 제시한다. 본 연구에서 제시하는 협업 프레임워크와 온톨로지 구축을 위한 시스템 공학적 방법론을 이용하여 전자상거래를 위한 정보시스템들의 운용상의 통합 기반을 마련할 수 있을 것이다. As the lifecycle of information technology becomes shorter and the characteristics of related personnel and projects become more diversified, there come to transformations of e-Business environments in these days. Moreover, the heterogeneous and distributed information of e-Business environments are not machine-understandable. In order to integrate the heterogeneous and distributed information of e-Business environments, self-describing protocols or integrative software design methodologies are needed and ontology would play an important role. In this study, the author proposes a framework for e-Business integration, which is based on a collaborative interoperable environment and the syntactic/semantic integration. The environment should be constructed by meta-modeling and employment of agents. The author classifies the agents by interface agents for message exchange, task agents for intra-process, and information agents for interconnection of legacy systems. The author is responsible for managing semantics by using the ontology and could make correct and automatic transactions between them. And the author uses DAML (DARPA Agent Markup Language) for specification of meta-models and builds ontology through the systematic approach. The ontology is represented in data and process meta-models, separately. The meta-models are built in a layered architecture based on the IRDS (Information Resource Dictionary Standard ISO/IEC 10027:1990) framework. The layers in this study are DAML layer, PSL-ontology layer, e-Business application layer, and XML Schema layer. The PSL-ontology layer is presented in modeling the process. An illustrative example of e-Business scenario is given, which is quoted from the RosettaNet standard and is concerning order management, such as quotation and purchase order. And the ontology is applied to the scenario. Finally, the methodologies for using and evaluating ontology are proposed, which are about the information retrieval, inferences, and evaluation criteria. And then, a prototype with database is implemented. By defining, sharing and managing resources and business-processes on the proposed ontology-based framework, it is possible to get rid of the cause of redundancy and inconsistency in e-Business environments.

Towards building an ontology repository

Maharjan, Roshan 경희대학교 2009 국내석사

The ontology repository has widely been used for storing and query-ing over ontology data. Recently the ontology repository gets more attentions as a component of ubiquitous systems. Lots of researches in ontology reposi-tories are focused on improving index algorithms, attaining more scalability, query answering, completeness and soundness of results of ontology data in context of ubiquitous systems. Since ontology repositories are different from conventional database sys-tems in the sense they are used for storing, managing and querying semi-structured data primarily ontology data. Therefore some index structures used for storing structured data in relational models, object relational models etc doesn’t perfectly fit for storing ontologies. In this thesis, we propose use of the path index to answer complex que-ries in ontology repository. Studies show that two types of queries namely chain queries and star queries frequently occur in SPARQL queries. Instead of using join operations over several triple indexes, we proposed to use path indexes for such queries, which in turn improve the query performance re-ducing the number of multiple join operations. Our experiments show that the proposed method enhances the query performance than using only index-es based on single triple pattern. Ultimately we use this combination of these indexes to build an ontology repository for improved query performance. Besides we present how we can achieve effective querying of web distri-buted ontology repositories. With OWL-Lite vocabulary and SPARQL pro-tocol on mind, we show how our method allows to have completeness of re-sults from proper study of reasoning completeness and careful formulation of triples from distributed repositories.

Enhancing Ontology Learning with Machine Learning and Natural Language Processing Techniques

Liu, Yue ProQuest Dissertations & Theses Rensselaer Polytec 2019 해외박사(DDOD)

소속기관이 구독 중이 아닌 경우 오후 4시부터 익일 오전 9시까지 원문보기가 가능합니다.

Ontologies provide terms to describe and represent specific knowledge and are thus widely used in many semantic web applications for knowledge management purposes. Since creating ontologies manually can be extremely labor-intensive and time-consuming, there is an increase in the motivation to automate the process, which includes the automation of ontology components such as classes, relations, attributes, and the overall term coverage and structure. Towards the goal of ontology automation in semantic web, in this thesis, we focus on the portion of ”Ontology Learning” that seeks automatic or semi-automatic approaches for either creating or reusing existing ontology resources with certain task-oriented objectives. We aim to enhance aspects of ontology learning with methods that can be reused in different domains and applied at large scale using machine learning and natural language processing techniques. We consider classes, relations, attributes being three major components of ontology and study three areas of ontology learning with respect to the automation of these components. In the first area, we study the extraction and linking of ontology classes from unstructured text, which aims to enhance the automation of ontology classes. In the second area, we study the knowledge extraction in a structured format from unstructured text, which aims to enhance the automation of ontology classes, relations, attributes as a whole. In the third area, we study the automatic ontology generation based on components like classes, relations, attributes we automatically learned from raw data. In this thesis, we contribute robust approaches with novel applications of Machine Learning and Natural Language Processing techniques to address three representative tasks in all three areas we studied that are closely related to the process of ontology automation. In the first contribution, in which we study the extraction and linking of ontology classes from unstructured text, we present a method for extracting noun phrases in the text that can be ontology classes and linking those to existing appropriate ontologies. Instead of defining new ontology terms, We demonstrate the method by linking the 3000 noun phrases extracted from biomedical and clinical literature to terms in existing biomedical ontologies with an innovative application of the word embedding technique. We demonstrate that training with a task-oriented resource can achieve state-of-the-art performance on a variety of tasks, such as abbreviation expansion and synonym detection, especially on domain-specific texts, such as clinical notes and an electronic health record (EHR). Our proposed approach has been adopted by many external researchers in the area and we are able to enrich over 1000 classes in 112 biomedical ontologies. In our second contribution, we study the knowledge extraction in a structured format from unstructured text. We focus on RDF triple extraction from unstructured text with the resulting triples being compatible with a given vocabulary, enabling knowledge to be represented in a machine-readable way. Inspired by neural machine translation, we propose an end-to-end model that extracts facts in the form of RDF triples directly from unstructured text. Our model outperforms the state-of-the-art pipeline based methods including entity linking and relation extraction on three different kinds of data sets. We demonstrate that the use of Knowledge Graph (KG) Embeddings in Machine Learning models have the potential to bridge the gap between natural language text and graph nodes, thus improving overall performance. Our third contribution to this thesis is a data-driven approach for bottom-up ontology generation with a specific objective of learning a harmonized catalog of items. In this task, we study a harmonized product catalog ontology in the area of eCommerce, where we have 730,901 e-Commerce product data scraped from the web. With an objective to improve categorization and faceted search, we propose effective approaches to build the ontology through a series of tasks including attribute extraction, product encoding and concept hierarchy generation with machine learning and natural language processing techniques. We demonstrate that the generated harmonized product catalog ontology, using our data-driven approach, can be used to make better product categorization and faceted search based on industry provided criteria.

(A) framework for ontology integration based on XMDR+

김정동 Graduate School, Korea University 2012 국내박사

Ontology integration carries significant importance in terms of reuse of knowledge and interoperability. Conceptually, ontology consists of two levels: schema and instance, each of which is designed and designed and developed by domain experts. Various distinct problems lie in ontology integration due to the matters concerning (i) how to express in which technical language, (ii) terminology, (iii) level of schema, (iv) level of instance and (v) system. To secure interoperability between metadata registries, an XMDR has been designed upon application of the ontology technology applied to semantic web. Despite the intent, the newly developed XMDR, which was to address the lack of the functionality in defining the relations among different MDRs, failed to accommodate all the ontological concepts. In other words, the conventional MDR carries problems concerning how to define and manage instant levels, since it mainly operates on the schema level of ontology. Besides, the XMDR does not address how to conduct mapping between ontological conceptual components and the MDR components needed for registration of ontology. Aware of the conditions, our study is to tackle the following two issues: (i) How to efficiently integrate ontology and (ii) why it is necessary to extended the existing XMDR. Then, we define an extended XMDR model titled “XMDR+ (eXtended Metadata Registry Plus),” and propose a new ontology-integrating framework that utilizes the XMDR+. As to the XMDR+, we have defined a model that accommodates diverse ontological relations such as ontological relations of schema and instant levels, properties and attributes. We further propose mapping rules, procedures and algorithms by defining mapping relations between MDR and the ontological components such that the latter are registered in the XMDR+ model. Last, the proposed XMDR+ model was designed and tested to demonstrate its capability of accommodating existing metadata registries (i.e. MDR and XMDR) and ontological concepts. To enhance the efficiency of the XMDR+ model-based integrated ontology framework, we qualitatively and quantitatively compared and evaluated traditional technologies and models for ontology integration. To sum up, the ontology integration framework proposed herein not only produces global ontologies; but also defines such integrating tools as mapping rules, procedures and algorithms, and, thereby, performs better in extending and integrating ontologies, than the other models like the Upper Ontology, the MOMIS and the LSD do. Moreover, our model successfully addresses the differences arising during integration, and, consequently, enhances reusability and interoperability.

Ontology-based Multi-Agent System for Supply Chain Management

Wang YongMing Kwangwoon University Graduate School 2009 국내석사

In the fast changing market environment, Supply Chain Management (SCM) has been considered as a powerful strategy to increase enterprise competency. This praceice enables enterprises to share of the information and cooperate even more closely. This paper explores the issue of how ontology-based Multi-Agent System (MAS) help to facilitate the communication among companies. Mas is expected to enable relevant agents participating team members from different companies to coordinate and communicate based on the roles and rules specified in the ontology. In this paper, Supply Chain Management System is composed of agent teams of different companies. Every agent team is able to cope with the ontology be used, when the members of Supply Chain is changed or some companie's system need to upgrade. The ontology in this paper is used to represent the service of agents and how to use these service..

지식의 공유 재이용을 위한 Group ontology에 관한 연구

장영일 朝鮮大學校 1996 국내석사

Very Large-scale knowledge based system orients the approach of knowledge sharing and reuse, where a frame of knowledge called common ontology is requested. Many ontology are being build, but each of them is implemented in first-order logic by using knowledge representation language. In such present circumstances of ontology, it occurs the situation such as top-down approach, exclusion of multiplicity and so on, that contradict sharing and reuse approach. This study proposes the concept-map system, the mechanism using visual information, which is a kind of technology, the environment of bottom-up knowledge representation by group. The core of the concept-map system is simplification of logical relations, where human knowledge is supposed to be made from special metaphor. And the concept-map itself allows logical expression described in semantic network. We examined how to specialize the logical expressions and how to manage the specialized knowledge, and did some experiments an that.

Domain Ontology Learning from Text Documents using Linguistic Patterns and Clustering Technique

이크발 카심 한양대학교 2013 국내박사

도메인 온톨로지는 개체들간의 개념적, 용어적 불합의 최소화를 통해 특정 도메인에 존재하는 개체들간의 의미적 불일치 (semantic gap) 문제를 해소하기 위하여 사용된다. 그러나 텍스트 문서로부터 지식 체계를 추출하고 도메인 온톨로지를 수동으로 구축하는 것은 도메인 전문가들에 의한 장시간의 협업이 필요한 매우 어려운 과정이다. 또한, 지식 체계 획득 및 온톨로지 구축을 자동화하고자 하는 연구 역시 구조화된 지식 표현 체계의 결핍 및 자연어 처리 기술의 한계와 같은 문제점으로 인하여 많은 어려움을 겪고 있다. 따라서, 도메인 온톨로지가 실제 환경에서 효과적으로 활용 가능한 도구로서 작용하기 위해서는 몇 가지 중요한 문제들이 반드시 극복되어야만 한다. 첫 번째 중요 문제는 텍스트 문서로부터 도메인 후보 용어를 획득하는 과정 및 이렇게 추출된 용어들 중 도메인 온톨로지 획득을 위한 주요 도메인 개념들을 효과적으로 획득하는 과정이다. 두 번째 중요 문제는 도메인에 존재하는 실제 지식 체계와 일치하는 도메인 개념들간의 분류적, 비분류적 의미 관계를 추출하는 과정이다. 최근, 도메인 온톨로지를 자동으로 혹은 반자동으로 구축하기 위한 다양한 연구들이 진행되고 있다. 그러나, 이러한 연구들은 특정 도메인에 대한 말뭉치 (corpus)에 극히 의존적이라는 점과 특정 문헌들로부터 도메인 말뭉치를 구성하기 위하여 전문가의 수작업이 필요하다는 점과 같은 여러 가지 한계점들을 가지고 있다. 또한, 기존의 연구들은 텍스트 문서에 존재할 수 있는 주요 명제들을 누락시킬 수 있는 대명사의 대용 해소 문제를 고려하지 않고 오직 명사 구만을 도메인 온톨로지 구축에 활용하고 있다는 문제점이 있다. 본 학위논문에서는 텍스트 문서로부터 지식체계를 추출하기 위한 새로운 방식의 도메인 온톨로지 구축 기법을 제안한다. 본 학위논문은 언어적 패턴 및 클러스터링 기법을 활용한 반자동적, 도메인 독립적, 비감독 기법을 제안함으로써 기존연구들과의 차별성을 보인다. 본 학위논문에서는 도메인 온톨로지 구축 과정을 위하여 다음과 같은 기법들을 개발하였다: 1)의미적 언어 패턴을 이용한 후보 용어의 추출 기법 및 의미적으로 연관성이 있는 개념들의 클러스터링을 통한 도메인 개념의 선정 기법, 2) Hearst 패턴을 활용한 도메인 개념간의 분류적 관계성 추출 기법, 3) 언어 패턴을 활용한 도메인 개념간의 비분류적 관계성의 추출, 명명 및 방향성 할당 기법, 4) 중간 단계 지식 체계 표현을 위한 도메인 개념도의 구축 기법, 5) 도메인 개념도로부터 도메인 온톨로지를 구축하기 위한 기법. 먼저, 언어적 타입 의존 규칙을 활용하여 텍스트 문서로부터 후보 용어들을 추출한다. 그 후, 추출된 후보 용어 쌍 간의 의미적, 구조적 유사도를 바탕으로 용어쌍 간의 최종 유사도를 계산한다. 이렇게 계산된 용어쌍 간의 유사도 값들은 친근도 전파 알고리즘의 입력값으로 활용된다. 친근도 전파 알고리즘은 고품질의 데이터 견본 (exemplar)가 발견되어질 때까지 데이터 포인트 간의 메시지 교환을 반복적으로 수행하는 데이터 클러스터링 알고리즘이다. 친근도 전파 알고리즘을 활용하여 추출된 모든 견본 데이터 포인트 (즉, 도메인 용어)들은 도메인 온톨로지를 학습하기 위한 도메인 개념으로서 활용된다. 그 후, 개념도 구축을 위하여 각각의 클러스터 내의 후보 용어들 간에 분류적/비분류적 관계성이 할당된다. 최종적으로, 구축된 개념도 개체들을 온톨로지 개체들로 변환하는 과정을 통하여 도메인 온톨로지를 획득한다. 마지막으로, 다양한 실험을 통하여 제안하는 시스템의 성능을 검증한다. 실험을 통하여, 본 학위논문에서 제안하는 기법을 통하여 구축된 도메인 온톨로지는 도메인 전문가의 수작업을 통하여 생성된 온톨로지와 일치함을 볼 수 있었으며, 도메인 전문가에 의한 평가를 통하여 구축된 도메인 온톨로지가 정보 시스템 도메인 및 학계 도메인의 지식 체계와 높은 수준으로 일치한다는 것을 확인할 수 있었다. Domain ontology can be used to bridge the semantic gap among the members of a domain through minimization of conceptual and terminological incompatibilities. Extracting knowledge from text documents and learning domain ontology manually is, however, a difficult, controversial, lengthy, and time consuming task that involves domain experts. Automatic or semi-automatic knowledge acquisition and ontology learning is also a non-trivial task due to the lack of structured knowledge representation and most of the data in documents are available in a free text format. Therefore several barriers must be overcome before domain ontology becomes a practical and useful tool. First important issue is the acquisition of candidate terms from text documents and then selection of domain concepts from these extracted terms for domain ontology learning. Second important issue is the extraction of taxonomic and non-taxonomic relationships between domain concepts which contain the actual context of a domain. Recently, various approaches for automatic or semi automatic construction of domain ontology have been proposed. However, these approaches suffer from several limitations such as heavy dependency on domain specific corpora and manual effort of experts required to populate the domain corpus from selected literature. Also, these approaches consider only the noun phrases in the text documents without resolving the anaphora resolution problems for pronouns which leads to miss the important propositions available in the text documents caused to decrease the recall. The proposed system presents a novel approach for domain ontology learning, defining new techniques for knowledge extraction from text documents. The utilization of linguistic patterns and clustering technique to the free text documents composing a semi-automatic, domain independent and unsupervised approach distinguishes the proposed system from the previous systems. We have been developed the following methods for the domain ontology construction process: 1) extraction of candidate terms using semantic linguistic patterns and selection of domain concepts by clustering semantically related concepts, 2) extraction of taxonomic relationships between domain concepts using Hearst’s patterns, 3) extraction, labeling, and assignment of direction of non-taxonomic relationships between domain concepts using linguistic patterns, 4) construction of domain concept map form extracted knowledge used as intermediate level knowledge representation, 5) finally, construction of domain ontology from constructed concept map. First, we extract candidate terms from documents using typed dependency linguistic rules. Second, Diset similarities are calculated based on semantic and structural similarity between pairs of extracted candidate terms. We then exploit affinity propagation algorithm, which takes as input Diset similarities between pairs of extracted candidate terms called data points. Real-valued messages are passed between candidate terms until a high quality set of exemplars iteratively emerges. All exemplars will be considered as domain concepts for learning domain ontologies. Then, extracted relationships are assigned between candidate terms in each cluster to complete the concept map. Finally, domain ontology is obtained from the constructed concept map by transforming concept map entities into domain ontology entities. The whole methodology has been implemented using different programming tools, providing a scalable solution. Finally, we verify the appropriateness of the proposed system by experimental results. Our empirical results show that the semi automatically constructed domain ontology conform to the outputs generated manually by domain experts, since the degree of difference between them is proportionally small. Also, domain experts have verified that the constructed domain ontologies are highly accordance with their knowledge and perception about information system domain and academia domain.

구조 정보 기반의 온톨로지 커널을 이용한 온톨로지 정렬

김성택 경북대학교 대학원 2011 국내석사

An ontology is a formalized specification for conceptualization in a specific domain. Thus, a computer can understand the human knowledge by adopting an ontology. However, since there exist numerous ontologies which are differently expressed with same human knowledge in the same domain, ontology interaction is too hard in the semantic web field. Ontology Alignment is one of the methods to solve the problem. Given two ontologies, ontology alignment aims to combine entities which have the same semantic. In this paper, we proposed a graph kernel specialized for ontology alignment. An ontology consists of five components such as concept, instance, data type, data value, and property. In the proposed kernel, ontology kernel, these components are considered when they are compared. Thus, the ontology kernel can be reflect more valuable structure information on an ontology than the ordinary graph kernel. The experimental results show that ontology kernel outperforms the ordinary graph kernel with respect to both the performance and practical computational time. In comparison with ontology alignment systems from OAEI 2009, the ontology kernel also get more advanced performance than others.