Earticle

An increasing number of English loanwords in Korean texts are one of the heavy obstacles to the automatic analysis of Korean texts due to the absence of these words in existing electronic lexicons. This study aims at proposing a semantic classification of 2,300 English loanwords that we selected from Korean Electronic Loanword Dictionary DECO-ZNF(Nam, 2010). The principle used in the semantic classification of Korean common nouns DSeC-N has been adopted in this study as well. On the basis of our results, we expect to predict in what semantic domains we should expand loanword entries: loanwords in semantic classes of new products, food, and art appear more important in number. The experiment presented at the end of the paper shows that a corpus of cosmetic products contains English loanwords much more than that of political news. Note that unless we expand the loanword entries related to this domain, the performance of automatic analysis will be far from being satisfying

Nowadays, Islamic content is widely used in research, including Hadith and the Al-Quran. Both are mostly used in the field of natural language processing, especially in text classification research. One of the difficulties in learning the Al-Quran is ambiguity, while the Al-Quran is used as the main source of Islamic law and the life guidance of a Muslim in the world. This research was proposed to relieve people in learning the Al-Quran. We proposed a word embedding feature-based on Tensor Space Model as feature extraction, which is used to reduce the ambiguity. Based on the experiment results and the analysis, we prove that the proposed method yields the best performance with the Hamming loss 0.10317.

The paper proposes that the Basic Concepts Classification (BCC) could serve as the controlled vocabulary for the Semantic Web. The BCC uses a synthetic approach among classes of things, relators, and properties. These are precisely the sort of concepts required by RDF triples. The BCC also addresses some of the syntactic needs of the Semantic Web. Others could be added to the BCC in a bottom-up process that carefully evaluates the costs, benefits, and best format for each rule considered.

전형적인 사역（강제사역）이라고 하는 것은 누군가가 누군가를 움직이게 해서 그 타인에 대해 무언가를 하도록 하는 것이나, 본고에서는 사역주가 무생물주어의 경우의 ｢(さ) せる｣ 구문의 의미에 관해서 고찰했다. 무생물주어의 구문에 있어서 ｢(さ) せる｣ 구문의 의미가「가능｣,「속성가능｣,「방법｣의 3가지의 의미를 나타내는 것을 밝혔다. 첫째, 주격명사구가 무생물의 의미를 나타내고 있는 것이 공통점이다. 또, 무생물은 ｢강제｣하는 것도, ｢허용｣하는 것도 불가능하다. 따라서 무생물주어에 있어서 양호한 조 건이 잘 갖추어져 있고, ｢무생물 주어는 ~をさせる｣는 가능의 의미가 된다고 생각되어진 다. 다시 말해서, 주격명사구라고 하는 것은 무생물에 내재되어 있는 속성(특성)을 서술 하는 문이 된다. 라고 하는 것이다. 무생물명사구에 관해서 속성을 묘사한다. 일종의 「속성 명사문｣이라고 하는 공통점이 있다고 생각되어진다. 둘째, 무생물주어의 사역문과 관련해서 선행연구에서는 넓은 의미의 인과관계로 파 악하는 것이 일반적이다. 그러나 본 논문에서는 인과관계의 외연이 너무나도 포괄적이 고, 속성명사구(무생물주어의 속성)이「가능｣의 의미를 갖는 구문이라고 분석했다. 무생 물주어에는 주격명사구가「속성을 서술｣하는 의미를 갖는 명사구가 상당히 존재하고, 그 주격명사구에는「속성｣을 나타내고 합해서「(さ) せる｣ 술어는「가능｣의 의미도 함께 포함하고 있다고 생각되어진다.「속성가능｣,「추상적 가능｣ 사이에, 2종류의 속성이 있 고, 은혜를 받는다고 하는 의미가 강하고, 마이너스적인 원인에 의한 인관관계가 아니 고, 더욱 상세하게 분류가 가능하다. 셋째, 무생물주어의 사역문에는「させる｣가「방법｣의 의미도 나타내고 있다고 생각되 어진다. 무생물주어의「(さ) せる｣가 특정의 목적을 위해 달성하려고 하는 ｢방법｣(수단) 으로서 사용되어진다고 생각한다. 그것은, ｢가능｣도 아니고, 인과관계도 아니다. 선행연 구에서의 인과관계의 외연이 너무나도 넓기 때문에 이 무생물주어의 수단으로서 취할 수 있다. 그렇기 때문에 이것은「방법｣의 의미를 나타낸다고 할 수 있다.

Typical causative (forced causative) is making others move by someone and thereby making them do something. But in this paper, the writer discussed the meaning of the sentence of ｢(さ)せる｣where causing subject is inanimate subject. In the sentence with inanimate subject, the meaning of ｢(さ)せる｣indicates ｢possibility｣ ｢possible attribute｣and ｢methods｣. First, it is common aspect that subjective noun phrase indicates the meaning of inanimate object. In addition, inanimate object can neither “force” nor “allow”. Accordingly, in case of inanimate subject, good conditions are prepared well and “inanimate subject を さ せ る ~ ” is thought to be indicating the meaning of possibility. In other words, subjective noun phrases are statements that describe properties (characteristics) that are inherent in inanimate objects. They describe properties about inanimate noun phrases. And they are thought to have common aspects of “attribute noun phrases”. Second, in the previous study related to causative sentences with inanimate subject, it is common to identify them as broad sense causal relationship. But in this paper, the writer analyzed that denotation of causal relationship is too comprehensive, and attribute noun phrases (attributes of inanimate subject) have the meaning of possibility. It is thought that in case of inanimate subject, there are many noun phrases where subjective noun phrases “describe attribute” and in such subjective noun phrases, predicate “(さ) せ る” also includes the meaning of “possibility”. Between “possible attribute” and “abstractly possible”, there are two attributes. The meaning of receiving grace is strong and it in not causal relationship caused by negative cause and more detailed classification is possible. Third, it is thought that in causative sentences with inanimate subject, “さ せ る” also indicates the meaning of “methods”. That is, inanimate subject’s “(さ) せ る” is used as “methods”(means) to be intended to achieve for specific purpose. It is neither “possibility” nor “causal relationship”. Because denotation of causal relationship in previous studies is too broad, it can be understood as means of this inanimate subject. Therefore, it can be said to be indicating the meaning of “methods”.

“게임 그래픽이 좋다.”라는 표현은 사람마다 해석을 다르게 할 수 있는 여지가 많다. 따라서 게임 그래픽 연구가 활발히 이뤄지기 위해서는 먼저 게임 그래픽을 체계적으로 분류할 필요가 있다. 본 연구는 게임 그래픽 세부 분류를 체계적으로 정리하는 것으로 게임 그래픽 연구에 대한 가이드라인을 제시해 게임 그래픽 연구 활성화에 기여하는데 그 목적이 있다. 게임 그래픽의 유형은 크게 2D와 3D로 분류된다. 게임 그래픽 구성요소는 원화, 캐릭터, 배경, 오브 젝트(프롭), 애니메이션, UI, 이펙트, 동영상 8가지로 분류할 수 있다. 게임 그래픽은 스타일, 퀄리티, 분위기로 세분화 할 수 있다. 북미는 3D 실사를 추구해 그래픽 퀄리티에서 높게 평가받고 있으며, 일본은 3D 실사, 카툰 렌더링, 애니 메이션풍을 게임마다 어울리게 적용해 그래픽 분위기에서 높게 평가받고 있으며, 유럽은 그래픽 퀄리티와 분위기에 통합적으로 투자하면서 빠르게 북미와 일본을 추격하고 있음을 확인했다. 도출된 모형을 게임 그래픽을 분석하는 툴 로 활용해 권역별 게임을 선정해 사례연구를 통해 분석했다.

The expression that “Game graphic is good.” can be interpreted differently by different people. Therefore, in order for game graphic research to be actively conducted, it is necessary to first systematically organize the detailed classification of game graphic. The purpose of this study is to systematically organize the detailed classification of game graphics and to provide guidelines for game graphics research, thereby contributing to the activation of game graphics research. The types of game graphic are largely classified into 2D and 3D. Game graphic components can be classified into 8 categories: original drawing, characters, backgrounds, objects(props), animations, UI, effects, and videos. Game graphic can be classified into style, quality, and atmosphere in meaning. North America is highly evaluated for its graphic quality by pursuing 3D photo-realism. Japan is highly evaluated for its graphic atmosphere by applying 3D photo-realism, cartoon rendering, anime style to games. And Europe is quickly catching up with North America and Japan by investing comprehensively in graphic quality and atmosphere. The derived model was used as a tool to analyze game graphics, and games were selected by region and analyzed through case studies.

최근 시맨틱 웹 문서는 클라우드 기반으로 생성 및 유통되고 문서유형 분류에 따른 쉽고 신속한 정보 검색을 위해 지능형 시맨틱 에이전트를 요구하고 있다. 기존의 웹 문서의 검색은 키워드를 이용하여 해당하는 질의어가 포함된 문서 목록을 결과로 가져오며 사용자의 요구시에 내용을 제시하는 것이 일반적인 형태이다. 이는 웹 문서의 유사도와 시맨틱 관련성을 고려하지 않음으로써 사용자가 내용 검색과 분석에 많은 시간과 노력을 요구한다. 이의 해결을 위해서 빅 데이터 요소 기술인 하둡과 NoSQL을 활용하여 시맨틱 웹 문서에 포함된 키워드 빈도에 기반한 웹 문서의 유형 분류와 유사도를 제시하는 시맨틱 클러스터링 모델을 제안한다. 제안 모델은 실시간 데이터 처리가 요청되는 이종 모델을 가진 공공 데이터와 웹 데이터를 취합하여 일반 사용자가 쉽게 질의할 수 있는 대용량 지식 기반 시스템을 구축하는데 응용 모델로 활용될 수 있다.

Recently semantic web document is produced and added in repository in a cloud computing environment and requires an intelligent semantic agent for analytical classification of documents and information retrieval. The traditional methods of information retrieval uses keyword for query and delivers a document list returned by the search. Users carry a heavy workload for examination of contents because a former method of the information retrieval don't provide a lot of semantic similarity information. To solve these problems, we suggest a key word frequency and concept matching based semantic clustering model using hadoop and NoSQL to improve classification accuracy of the similarity. Implementation of our suggested technique in a cloud computing environment offers the ability to classify and discover similar document with improved accuracy of the classification. This suggested model is expected to be use in the semantic web retrieval system construction that can make it more flexible in retrieving proper document.

Classification of Metaphors and Their Semantic Representation Young Hern Lee (Chosun University) The purpose of this paper is to represent the meaning of metaphors depending on their types. Miller(1979) classified the metaphorical expressions into nominal metaphor, predicative metaphor, and sentential metaphor, and represented their meanings such as nominal metaphor; Be(x,y) → (∃F)(∃G) {SIM [F(x), G(y)], and predicative metaphor; G(x) → (∃F)(∃y) [SIM [F(x), G(y)]] when a x is not G, and sentential metaphor; G(x) → (∃F)(∃y) [SIM [F(x), G(y)]] when y is not a discourse referent. However, those classifications of metaphors and semantic representations are not appropriate to formalize the metaphorical expressions and their meanings. First, Miller's classification does not cover the various types of metaphors. Second, the operator SIM is not necessary to describe the metaphorical meaning. Third, it is very hard to give the meaning of the argument forms of F(x) and G(y), In order to solve the problems above mentioned, we employ the subatomic semantics by Kearns(2000), which expended and revised Davidsons(1967). Every expression of metaphor can be decomposed into a subatomic formula, so that it is not necessary to define the meaning of F(x) or G(y), and to implement the abstract operator SIM. In addition to these, Complex Noun Phrase such as “river of time” can not be analyzed in Miller's mechanism because this complex noun phrase expression does not belong to the three types of classification that Miller(1979) analyzed. However, we can represent the metaphorical meaning of Complex NP more clearly. For instance, the expression “river of time” can be represented such as ∃e,e'[Pass(e) & Flow(e') & Theme(time,e) & Theme(river, e') & Imply(e, e')]. We also give the representations of the linguistic or metaphorical meaning of the expressions with evaluative and evidential adverbs.

This work concerns the ontological status of the common nouns of French and its sub-classification. We also propose a formulation of the sub-classification by the algebra of Boole.On the basis of the works of Lyons and Peterson, but also Vendler and Kiparsky & Kiparsky, we establish a criterion of the sub-classification : thing, event, fact, proposition.And so we propose following four subclasses:common nouns of thing : voiture(≒car), étudiant(≒student)common nouns of event : découverte(≒discovery), propagation(≒propagation)common nouns of fact : mystère(≒mystery), reconnaissance(≒recognition), conséquence(≒consequence)common nouns of proposition : supposition(≒supposition), souhait(≒wish), probabilité(≒probability), hypothèse(≒hypothesis)For each of the subclasses, we propose following four filtered boolean algebras:NCc = <{x : x ∈ P, x ≤ chose}, 1NCc, 0NCc, ∧NCc, ∨NCc, ’NCc> (where x’NCc = x’NC ∧ chose)NCe = <{x : x ∈ P, x ≤ événement}, 1NCe, 0NCe, ∧NCe, ∨NCe, ’NCe> (where x’NCe = x’NC ∧ événement) NCf = <{x : x ∈ P, x ≤ fait}, 1NCf, 0NCf, ∧NCf, ∨NCf, ’NCf> (where x’NCf = x’NC ∧ fait)NCp = <{x : x ∈ P, x ≤ proposition}, 1NCp, 0NCp, ∧NCp, ∨NCp, ’NCp> (where x’NCp = x’NC ∧ proposition)This formalization reflects the idea of the boolean semantics that all the possible categories possesses the boolean structure, and reconstitutes strictly the proposed sub-classification of the ontological point of view.

The image retrieval performance based on single feature is limited. For different kinds of images, it can not a better retrieval result. This paper raises image retrieval method based on weighted multi feature. In each kind of images, each feature precision is the weight evidence. On this basis, we research the existing semantic retrieval technology. Choosing the SVM classification theory which is more mature. Selecting parts of images as training set. Doing the training classification. From the research of different characteristics priority, it raises the image retrieval technology which synthesizes SVM and multi feature. From that, it can get a higher retrieval efficiency.

Semantic Web services have emerged as the solution to the need for automating several aspects related to service-oriented architectures, such as service discovery and composition, and they are realized by combining Semantic Web technologies and Web service standards. The increasing number of available Web services has raised the need for their automated and accurate classification. Services Classification would be useful to enact powerful capabilities such as automatic selection, invocation, composition, location or discovery of web services. In this paper, We proposes a approach of Semantic Web Services classification taking into account the feature of textual description and the semantic annotations of OWL-S advertisements. An experimental test of the proposed techniques is reported, showing positive results, the efficiency and accuracy of the automatic service portfolio has been improved.

The nature of the concepts regarding images in many domains are imprecise, and the interpretation of finding similar images is also ambiguous and subjective on the level of human perception. To solve these problems, in this paper, images’ semantic categories and the tolerance degree between them are defined systematically, and the approach of modeling tolerance relations between the semantic classes is proposed. Furthermore, for removing the induced false tolerance in the produce of using semantic tolerance relation model, the method of un-tolerating is introduced in image representation. We apply the proposed approach to the representations of images regarding the nature vs. man-made domain, human vs. nonhuman domain, and temporal domain, and compare the categorization results of them with the results not using semantic tolerance relation model. The results show the effectiveness of proposed method.

This paper mainly conducted two series of experiments. One is investigation between language dependent and language independent features. Bi-grams in Korean experiments and uni-grams in Chinese contributed most as basic features. And another one is utilization of Korean WordNet to improve the performance of Korean document classification. Korean WordNet is a Korean Lexical Semantic Network. Language independent features seem can lead better performance and stable. The performance of Korean text classification was improved by using Korean WordNet.

At present, information retrieval systems are simply expressed with a combination of keyword search according to the direct keyword matching method to get the information that users need. Because of this, documents retrieval systems serve too many documents due to term ambiguity. This makes the user need extra time and effort to get closer the document. To overcome these problems, this paper proposes the information retrieval system based on the content that connects documents according to the degree of semantic link that expresses a fuzzy value by fuzzy function. This paper also proposes an algorithm that produces a hierarchical structure using the degree of concept and content among documents. As a result, we are able to select and to provide user-interested documents.

In this paper, we surveyed newspaper editorials in both Japanese and Korean languages, and conducted a manual extraction survey of foreign words from editorials for 10 years (2013 to 2022). The samples obtained from this survey consisted of the number of editorials (Japanese) 6,592: (Korean) 8,977, the total number of characters) 1,555 words: (Korean) 1,035 words", the total number of words is “(Japanese) 39,590 words: (Korean) 26,167 words”. Using the top 100 words counted as high-frequency loanwords from this survey data, we classified the semantic fields based on the National Institute for Japanese Language and Linguistics (2004) “Classification Vocabulary Table (Enlarged and Revised Edition)”. First, the distribution of major classifications was common to both languages, with the noun “type of body” accounting for more than 90%. There were very few verb-related “用の類” and adjective-like “相の類”. However, based on the “Classification Vocabulary Table”, it is classified as “体の類”, but according to the research results of Kishimoto (2023) from a different perspective, some words are classified into nouns and verbs. touched. Words classified as verbs are “service, share, check, lead” in Japanese and “브리핑, 쇼핑, 인터뷰” in Korean. Next, in the middle classification, the distribution of “abstract relations” in Japanese and “human activity” in Korean was the most common. Furthermore, what is common to both languages is that six words overlap in the distribution of “language” in the “abstract relation” subclass. Among the subcategories, what is characteristic is that four words overlap in the field of “communications,” but there are also words related to the technology field. Korean) 빅데이터 was to be included. The results of the survey are a reflection of foreign- language products circulating at the time, but it remains to be seen how the results would change if the survey were to cover items other than editorials.

The ebXML (Electronic Business using eXtensible Markup Language) Specification Schema is to provide nominal set of specification elements necessary to specify a collaboration between business partners based on XML. As a part of ebXML Specification Schema, BPSS (Business Process Specification Schema) has been provided to support the direct specification of the set of elements required to configure a runtime system in order to execute a set of ebXML business transactions. The BPSS is available in two stand-alone representations, a UML version and an XML version. Due to the limitations of UML notations and XML syntax, however, current ebXML BPSS specification is insufficient to specify formal semantic constraints of modeling elements completely. In this study, we propose a classification schema for the BPSS semantic constraints and describe how to represent those semantic constraints formally using OCL (Object Constraint Language). As a way to verify a Business Process Specification (BPS) with the formal semantic constraint modeling, we suggest a rule-based approach to represent the formal constraints and to use the rule-based constraints specification to verify BPSs in a CLIPS prototype implementation.

이 글은 어촌 지역에서 다양한 의미로 실현되는 ‘-바리’를 그 의미 분화기제에 따라 의미 분류를 한 후 ‘-바리’의 결합 환경, 의미적 특징을 고찰해 ‘-바리’의 의미 파생 양상을 살펴본 것이다. 국립국어원의 민족생활어 사업에서 2010년부터 2012년까지 시행한 조사에 근거하여 동해안 지역의 강릉, 경주, 울릉도 어촌에서 조사된 생활어에 나타난 ‘-바리’형을 무작위로 추출하였다. 그 결과 ‘-바리’는 사전에 제시된 ‘어떤 성질을 많이 지니고 있는 사람이나 그 사람의 성질을 낮잡아 이르는 말’을 중심으로 ‘X를 잡는(채취하는) 행위 또는 X를 하는 사람, X를 잡기 위해 필요한 위치, 도구, 시기와 관련한 행위 또는 잡은 후에 하는 행위, X를 잡는 사람의 능력 또는 속성을 지닌 사람, X의 특정 부분 명칭, X 명칭’으로 의미 구분을 할 수 있었다. 의미 파생 양상은 다음과 같다. 첫째, ‘-바리’는 처음에는 ‘비바리’와 동일하게 어떠한 일을 전문적으로 하는 사람을 가리키는 말로 사용되다가 점차 그 사람이 갖는 속성과 관계되는 행위로 의미 확장이 일어나면서 지금의 ‘-잡이’의 의미와 유사한 의미를 지니게 된 것이다. 둘째, 어떤 대상물을 잡는 행위를 나타내는 ‘-바리’는 도구, 시기, 위치 등을 통해 의미적 파생이 일어났다. 이 중 시기는 하루를 기준으로 하면 ‘-치기’와 의미 속성을 공유하고, 절기를 기준으로 하면 ‘-지기’와 의미 속성을 공유하고 있다. 이는 사람을 나타내는 ‘-바리’가 행위로 1차적 의미 파생이 된 후 도구, 시기, 위치 등에 따라 2차 의미 파생이 된 것이다. 셋째, 해녀나 어부들이 하는 일에 대한 능력 또는 그런 속성을 지닌 사람을 ‘상바리, 중바리, 하바리’로 나누었는데 이는 해녀의 능력에 따라 ‘상군, 중군, 하군’으로 나누는 것과 동일한 방식과 의미로 파악되어 ‘-바리’의 의미가 중립적임을 알 수 있었다. 넷째, 단위명사인 ‘바리’와 접미사인 ‘-바리’가 공존하고 있음을 확인할 수 있었다. ‘머리, 마리> 바리’의 변화로 나타난 ‘-바리’는 ‘주변머리’와 같은 ‘비하’의 뜻을 더하는 접미사와 한 마리 또는 이천 마리를 의미하는 단위명사로 쓰여 두 개의 수량 단위로 사용된다. 다섯째, ‘다금발+-이’로 분석할 수 있는 가능성을 통해 물고기와 해조류 명칭에 나타나는 ‘-바리’는 접사 ‘-이’로 볼 수 있다. 이는 ‘-바리’와는 다른 어원에서 출발했을 가능성이 있다.

This study is to inquire into the semantic derivations of ‘-bari’ after examining the conjunctive circumstances and the semantic features of ‘-bari’. In doing so, this study classifies semantically ‘-bari’ according to its diversification mechanism, which is realized into various meaning in fishing villages. In order to achieve the goal of this study, based on the researches carried out from 2010 to 2012 in the national everyday word project of the National Institute of the Korean Language, this study extracts random samples of the ‘-bari’ types in the everyday words which were investigated in the fishing villages of the east coast areas such as Gangneung, Gyeongju, Ulleungdo. As a result, focused on ‘an underrating word for a man having a certain nature a lot and his nature’ presented in Korean language dictionary, it is found out that ‘-bari’ has following semantic diversifications: ‘an act to catch(collect) X or a man does X, an act in relation to location, tool, time which are necessary to catch X or an act after catching X, an ability of a man who catches X or a man who has the catcher’s attributes, a name of X’s specific part, X name ’. The aspects of the semantic derivations are as follows. Firstly, ‘-bari’ at first equally with ‘Bibari’ was used as a word which refers to a man who does a certain work professionally. And, with the gradual semantic extension into an act related with his attributes, ‘-bari’ comes to have a similar meaning with the today’s meaning of ‘-jabi’. Secondly, ‘-bari’ which refers to an act to catch a certain object underwent a semantic derivation by tool, time, location. Among them, the time shares semantic component in relation to ‘-chigi’ by the one day basis and the lunar calendar's marking of seasons shares semantic component in relation to ‘-jigi’. This means that ‘-bari’ which refers to a man underwent first semantic derivation by an action, and then underwent second semantic derivation by tool, time, location. Thirdly, a man who has some abilities or attributes for the work of Haenyeo(woman diver) or fisherman was diversified into ‘Sangbari, Jungbari, Habari’, which is the same method and meaning with the Haenyeo’s diversification into ‘Sanggun, Junggun, Hagun’ by the ability of Haenyeo. Fourthly, it is found out that unit noun, ‘Bari’ and suffix, ‘-bari’ coexist. The ‘-bari’ appeared by the change of ‘Meori, Mari>Bari’ is used as suffix to add the meaning of ‘belittlement’ like ‘Jubyunmeori’. The unit noun, ‘Bari’ was used as two units of the quantities meaning one fish or two thousand fish. Fifthly, The ‘-bari’ appeared in the name of fish and seaweed, through the possibility that it can be analyzed by ‘Dageumbal+-yi’, can be considered as affix ‘-yi’, which also suggests a possibility that its origin is different from that of ‘-bari’.

In Korean terminologies, adjectival terms are rare, and the meaning and function associated with adjectives in Indo-European languages are often realized instead in noun form. However, the rarer adjectival terms we, the more they are used in restrictive and repetitive ways in specialized domains. Thus, it is important to distinguish the different senses of these terms. In this work, focusing on semantic modeling in terminology, we distinguish the different senses of adjectival medical terms by applying lexico-semantic criteria (L'Homme, 2004a) and by classifying the semantic category of the arguments of the adjective (Bae and others, 2002). The result not only contributes to enriching medical terminology, but also empirically demonstrates a method for distinguishing the different senses of adjectival medical terms. In this work, we obtained an average of 1.854 senses for each term. We used the KAIST corpus, composed of medical texts (1,500,000 eojeols), and a group of texts on various subjects (40,000,000 eojeols)

본 논문에서는 온톨로지의 개념구조를 이용한 웹페이지의 의미적 분류방법을 제안한다. 웹 문서들이 가지는 용어 정보들과 어휘들 간의 개념 구조를 파악하여 온톨로지를 확장시키면서 이를 문서분류에 적용하여 의미적 분류가 이루어지게 한다. 문서 분류는 문서들을 가장 잘 표현할 수 있는 자질들을 정하고 이러한 자질들을 통해 미리 정의된 2개 이상의 카테고리에 문서의 내용을 파악하여 가장 관련이 있는 카테고리로 할당하는 것이다. 본 논문에서는 웹 문서에서 추출한 용어 정보들의 유사도와 온톨로지 카테고리의 유사도를 계산하여 웹 문서를 분류하여 문서 분류를 위한 실험데이터나 학습과정 없이 바로 실시간으로 문서분류가 이루어지며, 결과적으로 온톨로지와 문서들이 가지는 고유한 의미와 관계의 식별을 통하여 보다 더 정확하게 문서분류를 가능하게 해준다.