문장의 의미 해석을 위한 의미영역의 역할 - ≪상거래≫를 중심으로 -

이동혁 한국어의미학회 2008 한국어 의미학 Vol.26 No.-

The purpose of this paper is to take a closer look at the function of a semantic domain for interpreting a sentence meaning. A sentence meaning is understood through the common background that a speaker and a hear hold in, which the common background is the semantic domain. In order to interpret a sentence meaning, it is necessary to solve a several matters; judgement of acceptance degree, interpretation of structural meaning, and disambiguation of word polysemy. A semantic domain is involved in analyzing above matters. And a semantic domain is made up of domain elements, and it makes a network between domains, or between domain and ontological category. Domain elements have prototypical words, which are a criterion for judging a degree of acceptance of a sentence meaning. A structural meaning of the sa-ta sentence must be analyzed with projecting an ontological category LOCOMOTION and a semantic domain COMMERCIAL. Sa-ta has a lot of meanings, and its polysemy is able to be disambiguated based on a semantic domain. If several semantic domains are projected on a sentence, it should be interpreted figuratively. Semantic domains are corresponded through the generic space as an ontological category.

상징어의 의미영역 분류 연구

김홍범 ( Hong-beom Kim ) 한국문법교육학회 2016 문법 교육 Vol.26 No.-

The purpose of this paper is to classify Korean symbolic words according to the semantic domain. The process and methodology of this research is as follows; 1) Comparing the definitions and their related information of symbolic words in the various dictionaries. 2) Constructing an Excel database by dismantling Classification Dictionary of Korean Onomatopoeic and Mimetic Words . 3) Analyzing Standard Korean Great Dictionary and obtaining the material such as North Korean language, dialects, errata, and lists of similar type words. 4) Proposing a semantic description model to implement the consistent methodology of meaning clarification. Focusing on the problems of semantic-domain-based classification of symbolic words proposed by the existing dictionaries, major domains are reclassified as follows; 1) Subclassifying 930 ``symbolic words related to the solid`` out of 2116 onomatopoeic words. 2) Dividing ``sense`` into ``sense``, ``emotion``, and ``perception`` and establishing 8 upper categories for 40 domains of ``character``. 3) Reclassifying the ``movement`` domain containing 1,650 symbolic words out of 2,639 mimetic words. 4) Establishing domain layers for ``speaking shape`` in the sub-layer of movement. The major limitation of current classification system is the lack of understanding about top-bottom concept. The discrimination of symbolic words that belong to the same semantic domain is continuously required.

Semantic Domain Network Analysis of Maritime English Near-Synonyms

Se-Eun Jhang,Sung-Min Lee,Tony McEnery,Vaclav Brezina,Wenyu Lu 한국코퍼스언어학회 2017 Corpus Linguistics Research Vol.3 No.-

The purpose of this paper is to graphically represent semantic domain networks of maritime English near-synonyms as such displays can provide insights into semantic relations. A specialised corpus from international maritime communities was used. Keywords are extracted by comparing the Maritime English Corpus with the BNC Baby. Among keywords within the top 20 ranks, we focus on ship-vessel and maritime-marine. We use the MI3 score to identify collocates of the two pairs of near-synonyms and the Wmatrix web interface program to tag semantic domains. We create key semantic domain networks using the social network analysis tool NetMiner 4.0. The results show that some semantic domains are connected between the two pairs of near-synonyms, while others are not connected. The findings of the study indicate how semantic domain networks of keyword collocates and simple collocates help to distinguish near-synonyms through graphical representations of a corpus. The paper found that semantic domains with keyword collocates have a stronger tendency to connect two pairs of near-synonyms than those with simple collocates.

派生接辭 ‘-이’의 意味成分 分析

최상진 한국어문교육연구회 2011 어문연구(語文硏究) Vol.39 No.4

This paper examines the lexical meaning which the derivative affix '(-이)' comes to have with word roots in semantic formation of derivative words, and further, how its semantic domain and semantic features can be analyzed. What is to be specifically analyzed are derivative sub- stantial nouns for real things from which we are able to grasp their connotations and denotative concepts. Depending on combinative aspects of word roots or sentences as semantic components, derivative affix ‘-이’ comes to be able to cover a wide range of semantic domains. The semantic formation of derivative words consists, primarily, of the cooccurrence relation to arise between word roots and affix words; and, secondarily when affix words intervene as sentence components, it is constituted by a self-organization derived from its close interrelationship with sentence components. Semantic features of affix ‘-이’ can be analyzed separately between their connotative and denotative dimensions. Connotative semantic features can be analyzed in view of its semantic domain, whereas denotative ones are possible to be analyzed when derivative words are used as sentence components. 이 論文은 派生語 意味形成에서 派生接辭 ‘-이’의 語彙的 意味를 살펴보고, 그 意味領域과 意味資質은 어떻게 分析될 수 있는가를 考察한 것이다. 派生接辭 ‘-이’는 動詞 語幹에 접미되어 [+有情性]을 지닌 人物, 動物 또는 [-有情性]을 지닌 事物 등의 의미가 添加되어 特定한 個體를 지닌 實體名詞를 파생시킨다. ‘-이’는 語基 혹은 문장의 意味構成要素로서의 결합 양상에 따라서 다양한 意味領域을 가지게 된다. 파생어 의미는 1차적으로는 語基와 接辭의 相補的인 意味的 共起關係로 이루어지고, 派生語가 문장의 構成要素로 參與하면 2차적으로 文章構成素와의 相互間 有機體的 自己組織(self-organization)에 의해서 이루어진다. 파생접사 ‘-이’의 意味資質은 內包的 意味資質과 外延的 意味資質로 나누어 분석될 수 있다. 內包的 意味資質은 意味領域을 考慮하여 분석될 수 있으며, 外延的 의미자질은 파생어가 문장 속 構成要素로 쓰이고 있을 때 분석이 가능하다.

韩汉身体词“배”与“腹/肚”的语义延伸分析

LI NAN,한용수 중국문화연구학회 2018 중국문화연구 Vol.0 No.41

The abdomen belongs to a part of the body of the human body, and its related body-part terms can be regarded as “visible, felt” in vitro words; on the other hand, the abdomen includes abdominal wall, abdominal cavity and visceral organs, and it is related to the body words that are “invisible and unable to touch”. Therefore, the body-part terms related to the abdomen are both concrete and abstract. Heine et al.(1991)points out: “the conceptual domain has certain directivity.” This is the shift from the concrete “person” domain to the abstract “quality” domain. On the basis of the conceptual domain of Heine, the semantic extension of the Korean and Chinese body-part terms, “abdomen / belly”, is investigated in the order of person > animal > object > space > quantity > abstract concept domain. Its main contents are as follows. The semantic extension of the Korean body-part term “abdomen” is from the process of person > animal > object > space > quantity > abstract. The Chinese “abdomen” is in accordance with the process of person > animal > object > space > abstract process, and the extension meaning of Korean body word “abdomen” and Chinese body word “abdomen” is greater than difference. In addition to the meaning of “animal production times”, the extended meaning of Korean “abdomen” can be found in Chinese body words containing “abdomen”. However, the semantic extension of Chinese “abdomen” is more abundant in person and abstract domains. In the person domain, the “abdomen” can refer to “the protruding part of the finger or the calf”, and the “abdomen” can be combined with the “heart” for the person that can be trusted because of its functional importance, and the “belly” is able to borrow a pregnant person and a protruding person in the abdomen because of its appearance. In the spatial domain, the position of the abdomen in the body can also refer to “the front” and “the center”. In the process of semantic extension of Korean and Chinese body-part term, the main source of image is its structure, form, position and function. Structural features are more semantically transferred in the human domain itself based on metonymic mechanism of the whole generation. The direct perception of the morphological features is the strongest. It is the most easy to leave the mental image of the form and picture, and it is also the most likely to attract the attention of human to the appearance of other things. But there is no further meaning in terms of morphological characteristics. This is because the features of the morphology are characterized by a very high representational character. The more concrete, the greater the limitation of the development of the word meaning. “position” features like “morphologies” features, with strong intuitive perception, but “positional” features reflect a higher level of abstraction than “morphologies”. The expansion of the abdomen to the spatial domain is precisely because “location” is the spatial relationship between objects. The human abdomen is an important part of the trunk, including the stomach, the intestines, the kidney, the uterus and other visceral organs. Therefore, the human abdomen has the functions of digestion, excretion, and pregnancy, and the abdomen is closely related to the other parts of the body, such as “heart”, to develop into a more advanced abstract domain during the cognitive process. Extend the meaning of heart, idea, emotion, and so on.

한국어 연결어미 의미 확장에서의 환유와 은유

박재연 ( Bak Jae-yeon ) 국어학회 2014 국어학 Vol.70 No.-

이 논문에서는 한국어 연결어미에서 발견되는 의미 확장을 환유적 확장과 은유적 확장으로 나누어 살펴보았다. ‘선행’의 의미를 가지는 ‘-어서, -다가’ 등이 ‘원인’ 혹은 ‘조건’의 의미로 발달하는 의미 확장은 문맥에서 발생한 함축이 의미 성분으로 관습화하는 의미 성분 추가 현상에 의한 것으로서 인접한 의미로의 전이라는 점에서 기본적으로 환유적 성격을 가진다. 반면 ‘목적’의 의미를 가지는 ‘-느라고, -답시고’가 ‘원인’의 의미로 발달하는 의미 확장은 기존의 의미 성분의 일부가 삭제되는 의미 성분 삭제 현상에 의한 것으로서 역시 인접한 의미로 전이되는 환유적 과정이다. 그러나 연결어미의 의미 확장에는 이와는 전혀 다른 성격의 은유적 확장이 존재한다. ‘대립’의 ‘-지만’, ‘조건’의 ‘-으면’, ‘원인’의 ‘-으니까, -어서’ 등은 일반적으로 ‘내용 영역에서의 대립, 조건, 원인’을 나타내지만 ‘인식 영역에서의 대립, 조건 원인’이나 ‘화행 영역에서의 대립, 조건, 원인’을 나타내기도 한다. 이는 서로 다른 영역을 넘나들며 발생하는 은유적 과정이다. The aim of this paper is to differentiate two kinds of semantic extensions in Korean connective endings: one is metonymic and the other is metaphoric. The semantic extension from ‘preceding’ to ‘cause’ or ‘condition’ found in ‘-eoseo, -daga’ is the metonymic process by adding a semantic component. The semantic extension from ‘goal’ to ‘cause’ found in ‘-neurago’ and ‘-dapsigo’ is also the metonymic process by deleting a semantic component. By contrast, the metaphorical semantic extension can be found in connective endings. The primary senses of ‘-jiman, -eumyeon, -eunikka, -eoseo’ are ‘contrast, condition, cause’ in content domain, but they can be used as ‘contrast, condition, cause’ in epistemic domain or speech act domain. This semantic extension is a kind of metaphoric process by cross-domain mapping in the different two domains.

WSDL 기반의 시맨틱 웹 서비스 온톨로지 구축 방법에 관한 연구

강수호,이현정,손미애 한국기업경영학회 2010 기업경영연구 Vol.17 No.2

본 연구에서는 웹 상에 존재하는 웹 서비스들을 자동적으로 발견, 조합 및 실행하고자 하는 시맨틱 웹 서비스(semantic web service)를 구현하는데 필요한 온톨로지를 자동적으로 생성할 수 있는 알고리즘을 제안한다. 현재까지는 시맨틱 웹 서비스 구현을 위한 온톨로지가 지식 공학자에 의해 구현되었다면 본 연구에서 제안하는 알고리즘은 지식공학자의 온톨로지 구현을 보다 쉽게 가능하게 할 수 있는데 의미가 있다. 제안되는 알고리즘으로 생성되는 온톨로지는 골격 온톨로지(Skeleton ontology), 서비스 온톨로지(Service ontology) 및 도메인 온톨로지(Domain ontology)로 구성된다. 골격 온톨로지는 시맨틱 웹 서비스 구현을 위한 기본 온톨로지로 탐색된 WSDL 도큐먼트들 중 핵심 클래스를 가장 많이 포함하고 있는 문서를 이용해 구성된다. 서비스 온톨로지는 골격 온톨로지의 생성에 활용한 ‘base-WSDL 도큐먼트’를 제외한 나머지 WSDL 도큐먼트들을 이용해서 골격 온톨로지를 확장함으로써 생성한다. 이때 클래스들 간의 ‘주체-객체관계(subject-object relation)’ 및 인스탄스들이 정의되고 고려된다. 도메인 온톨로지는 모든 WSDL 도큐먼트들에 서비스 온톨로지가 적용되었을 때 생성 완성된 온톨로지이다. 본 연구의 우수성 검증을 위해 ‘Semantic Service Selection(http://www-ags.dfki.uni-sb.de/~klusch/s3/index.html)’에 구축되어 있는 ‘카메라 온톨로지’의 참조 온톨로지와 본 논문에서 제안한 알고리즘을 이용해 자동적으로 구축한 온톨로지 간의 유사도를 클래스들과 ‘주체-객체관계’ 관점에서 ‘Hit ratio’와 ‘Precision’을 평가하였다. 온톨로지를 본 연구에서 제공하는 알고리즘을 이용해 자동 구축하였을 경우 평균 ‘Hit ratio’는 57.2%, 평균 ‘Precision’은 81.5% 이다. 본 알고리즘이 지식공학자의 온톨로지 생성에 큰 기여를 할 수 있을 것으로 기대된다. In this study, we suggest WSDL-based Ontology construction algorithm for semantic web service to maintain web service on the Web. The algorithm is composed of three parts. First, this is to build of basic skeleton of Ontology in the specific domain. It is named by Skeleton Ontology. The algorithm uses a WSDL document (which is called by ‘base-WSDL ontology’) to build the Skeleton Ontology. Second, this part supports extension of the Ontology using the Skeleton Ontology. To abundant the Ontology, the algorithm extract classes and properties from all of collected WSDL documents except the ‘base-WSDL ontology’in the domain and add those to the Skeleton Ontology. It is called by Service Ontology. Finally, when applying of Service Ontology to whole of WSDL documents is completed, it is called by Service Ontology. To evaluate the performance of the WSDL-based Ontology construction algorithm, the algorithm is applied to camera domain and the newly generated ontology is compared with the existing reference ontology using Hit ratio and Precision. The result of Hit ratio is 57.2% and Precision is 81.5%. We conclude that the WSDL-based Ontology construction algorithmclearly discriminate the classes and subject-object relations.

“V来V去”格式语义特征小考

초팽염 중국학연구회 2017 중국학연구 Vol.- No.81

This paper tries to make some new explorations on the semantic features of ‘V+lai+V+qu’. Previous studies classified the meaning according to whether it was moved or repetition, but this article classified it into three cognitive domains according to the context where the ‘V+lai+V+qu’ structure is used. We found that the structure of ‘V+lai+V+qu’ has different semantic features in the content domain, epistemic domain, and speech-act domain. The difference in these features arises from the conceptual elements that are activated in the conceptual frame of "reciprocating behavior schema". On the other hand, from a pragmatics point of view, the ‘V+lai+V+qu’ structure has some additional meaning besides its core meaning. Since the structure of ‘V+lai+V+qu’ shows a larger amount of action than expected, it shows a lot of negative meaning in the content domain. In the logical domain, it has a negative or positive meaning depending on whether it meets the expected outcome. And in the speech-act domain we can also possible to confirm the strong subjectivity because it is a conclusion from comprehensive judgment. Through the three domains and the different perspective we can clarify the meaning of ‘V+lai+V+qu’ more clearly and deeply.

Domain-Adaptation Technique for Semantic Role Labeling with Structural Learning

임수종,이창기,류법모,김현기,박상규,나동열 한국전자통신연구원 2014 ETRI Journal Vol.36 No.3

Semantic role labeling (SRL) is a task in natural- language processing with the aim of detecting predicates in the text, choosing their correct senses, identifying their associated arguments, and predicting the semantic roles of the arguments. Developing a high-performance SRL system for a domain requires manually annotated training data of large size in the same domain. However, such SRL training data of sufficient size is available only for a few domains. Constructing SRL training data for a new domain is very expensive. Therefore, domain adaptation in SRL can be regarded as an important problem. In this paper, we show that domain adaptation for SRL systems can achieve state-of-the-art performance when based on structural learning and exploiting a prior model approach. We provide experimental results with three different target domains showing that our method is effective even if training data of small size is available for the target domains. According to experimentations, our proposed method outperforms those of other research works by about 2% to 5% in F-score.

Word Multi-Domain Semantic Polarity Algorithm Based on Domain Standard Word Set

Yun Sha,Shibo Zhang 보안공학연구지원센터 2015 International Journal of Hybrid Information Techno Vol.8 No.9

Text orientation is one of the most important base work in web semantic calculation. The semantic polarity of words is the basis element of text sentiment. But in traditional word orientation algorithm, each word gets orientation value based on its distance with these words in the standard word list. So the word polarity is always the same value. But in fact, the word may express different polarity in diverse domains. In this paper, word polarity is calculated based on domain standard word list according to its part of speech. The domain standard list is established by search engine result. The experimental result shows that the domain standard word set for word orientation can increase the accuracy rate of comments orientation calculate.