• Annual Conference on Human and Language Technology
• /
• 1993.10a
• /
• pp.461-468
• /
• 1993

이광오[1]의 연구를 바탕으로 한글 필기 행동의 설명을 시도하였다. 한글 필기에 나타나는 실수에는 예기, 지속, 부가, 생략 등의 유형이 있었으며, 실수에 개입되는 단위로는 단어, 글자, 자모가 인정되었다. 필기에 필요한 기본단위로 글자표상을 가정하고, 심성어휘집에서의 부분적 표기정보 인출과 음소-자모 변환에 의해 글자표상이 생성되는 필기행동 모형에 대해 고찰하였다.

• Journal of Information Management
• /
• v.21 no.1
• /
• pp.53-75
• /
• 1990

In information storage and retrieval system, controlled vocabularies are generally used to improve the recall ratio and to guide for indexers/users to select correct indexing/searching terms by regulating their forms as well as meanings. Thesaurs, a type of controlled vocabulary, nowdays accept ed by most of database producers. The objective of this study is to develope a method of Korean Language Thesaurus construction. This study covers 1) the definition of thesaurus, 2) a literatural survey on term relations and thesaurus construction method, 3) a suggestion for a practical construction method, 4) the display format of thesuari, and 5) tests and the results.

• Proceedings of the Korean Society for Cognitive Science Conference
• /
• 2002.05a
• /
• pp.151-156
• /
• 2002

중의적인 단어를 처리하는 방법에 대한 선행연구로, 첫째 문맥에 맞는 의미가 먼저 활성화된다는 가설과 둘째, 여러 뜻 중에 상대적인 빈도에 따라 많이 쓰이는 의미가 먼저 활성화되고, 그것이 문맥과 일치하지 않는다면, 다른 관련된 의미를 찾는다는 가설이 제기되었다. 마지막으로 문맥에 상관없이 모든 의미가 활성화 된 후 문맥을 고려하여 문맥에 적절한 의미를 선택한다는 가설이 있다. 본 연구에서는 '먹을', '감을' 등과 같이 2가지 의미의 품사가 다른 중의 어절과 '쥐어', '감어' 등과 같이 어절 문맥('어')이 주어진 어절의 의미 활성화가 어떻게 다른지를 조사하였다. 본 연구의 목적을 위해 점화어휘 판단 과제를 사용하였다. 실험 1의 결과는 SOA 150ms 조건에서 점화자극어절과 관련된 의미가 품사와 관련 없이 모두 활성화되었다. SOA 1000ms 조건에서는 상대적으로 많이 쓰이는 체언의 의미는 계속 활성화 되어 있는 반면, 용언의 의미 점화량은 감소하였다. 명칭성 실어증 환자인 SDK의 경우 SOA 150ms 조건에서는 일반인과 같은 형태소 처리특성을 보였으나 1000ms 조건에서는 달랐다. 다른 명칭성 실어증 환자인 BIS과 전반성 실어증 환자인 PSB는 SOA 150ms 조건과 1000ms 조건에서 일반인과 아주 다른 양상을 보였다. 이것은 실어증 환자의 타잎에 따라 형태소의 처리나 중의적인 의미 활성화가 일반인과는 다르다는 것을 보여준다. 실험 2에서는 어절 문맥이 있는 '먹어', '쥐어', '감어' 등과 같은 어절을 사용하였다. 실험 2의 결과는 SOA 150ms 조건일 때 어절문맥의 영향으로 용언의 의미만 촉진적 점화효과가 있었고, 체언의 의미는 활성화되지 않았다. 그러나 SOA 1000ms로 지연시켰을 때는 용언뿐만 아니라 체언의 의미도 촉진적 점화효과가 있었다. 실험 1과 2의 결과는 중의적인 한국어 어절의 경우에도 모든 의미가 활성화되나 어절 문맥이 존재할 때는 어절 문맥의 제약으로 어절 문맥에 맞는 한 가지 의미만 활성화된다는 것을 암시한다. 또한 이러한 결과는 한국어 어절이 분석된 형태가 아닌 어절 형태로 심성 어휘집에 저장되어 있다는 것을 암시한다. 실어증 환자의 경우 실험 1과 마찬가지로 환자의 수준이나 종류에 따라 다양한 반응을 보여주었다.

• Sijohaknonchong
• /
• v.21
• /
• pp.47-76
• /
• 2004

Newly discovered ${\lceil}Gogeummyeongjakga{\rfloor}$ is a copy collection of the old poetry SI-JO, which is specified in the book of yellow Gojeongji. {$\lceil}Gogeummyeongjakga{\rfloor}$ is guessed to be copied before the 17th year of King Yeongjo's reign(1740) and thus it's the early collection in the history of the Korean verse, Shijo. According to our research, there are 78poems in the collection and nine out of them hasn't been yet reported to the Korean Academy. The characteristics of Shijo in the book are followed. First. The collection is different from other books since the book was written in Korean instead of Chinese characters, which shows the uniqueness of the Korean literature in the late 17th and the early 18th century. Secondly, there are different versions of a poem in the collection, which is quite unusual in the other collections. There are different words or phrases used in different versions and even the whole verse is modified in some cases. Thirdly. two out of newly discovered nine short lyric songs is transformed from and that are kind of Chinese Ak-Bu. By the way, the compiler of ${\lceil}Gogeummyeongjakga{\rfloor}$ seemed to understand the co-relation between Ak-Bu and Shijo. and that's why he chose transformed Shijo from Ak-Bu not Chinese poetry. Among nine poems, <9> and <10> are newly discovered responding songs unknown up till now.

• Journal of the Korean Society of Costume
• /
• v.57 no.5 s.114
• /
• pp.140-150
• /
• 2007

The study aimed to classify the synonyms of clothing terms, such as equal relationship, connoting relationship and partial relationship. The subjects of the study are eleven Chinese character study books such as <石峰千字文 Seokbongcheonjamun> <丙子本千字文 Byeongjacheonjamun>, <註解千字文 Juhaecheonjamun>, <訓蒙字會 Hungmongjahoi>, <新增類合 Sinjeungyuhap>, <兒學編 Aahakpyeon(1816) (1908)>, <蒙喩編 Mongyupyeon>, <字類註釋 Jalyujuseok>, <正蒙類語 Joengmongyueo> and <通學徑編 Tonghakgyeongpyeon>, seven foreign language study books such as <譯語類解 Yeokeoyuhae>, <譯語類解補編 Yeokeoyuhae-supplementary book>, <同文類解 Dongnunyuhae>, <蒙語類解 Mongeoyuhae>, <蒙語類解補編 Mongeoyuhae-supplementary book>, <倭語類解 Oaeeoyuhae> and <方言類釋 Bangeonyuseok>, and some vocabulary books <才物譜 Jaemulbo>, <廣才物譜 Kwangjaemulbo>, <物譜 Mulbo>, <物名攷 Mulmyeongko>, and <事類博解 Salyupakhae>. There are two kinds of synonyms with equal relationship that are spelled as more than two names for the same clothing. The first group was names that have different vocabularies for the same Chinese character headword, for example, Gyeokji and Namosin, Jeoguriot and Dongdolssi, Deungjige and Got dongot, Daloi and Nangja, Jitbidan and Oaedan, jusa and jeuusya, jusa and Murui, Muja and Heoja, and so on. The second group was names that are spelled with sound and meaning of Chinese character, for example, Nuyeok and Saui, Binhyeo and Jam, Mosi and Jyeopo, and so on. Also synonyms of Chinese names were found in equal relationship. Synonyms with connoting relationship in which one name perfectly connotes meaning of another name are Gotgal and Susik, Danryeong and Gwandae, Bosyeon and Chyeong, ete. In these cases, the range of the meaning of clothing terms can be confirmed since names with broad meaning connote names with narrow meaning. There are differences in use and form. In partial relationship synonym with the same name, some synonyms like Baji and Goui, and jeoksam and Hansam have difference in use, while Gamto, Got and Sarno, Neolku and Satgat, and Nangja and Cheopji have difference in form. These presented differences among similar names.

• Journal of KIISE:Software and Applications
• /
• v.36 no.1
• /
• pp.82-91
• /
• 2009

Sign language spotting is the task of detecting and recognizing the signs in a signed utterance. The difficulty of sign language spotting is that the occurrences of signs vary in both motion and shape. Moreover, the signs appear within a continuous gesture stream, interspersed with transitional movements between signs in a vocabulary and non-sign patterns(which include out-of-vocabulary signs, epentheses, and other movements that do not correspond to signs). In this paper, a novel method for designing a threshold model in a conditional random field(CRF) model is proposed. The proposed model performs an adaptive threshold for distinguishing between signs in the vocabulary and non-sign patterns. A hand appearance-based sign verification method, a short-sign detector, and a subsign reasoning method are included to further improve sign language spotting accuracy. Experimental results show that the proposed method can detect signs from continuous data with an 88% spotting rate and can recognize signs from isolated data with a 94% recognition rate, versus 74% and 90% respectively for CRFs without a threshold model, short-sign detector, subsign reasoning, and hand appearance-based sign verification.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.8 no.4
• /
• pp.803-809
• /
• 2007

This study suggests a computational model to inquire the roles of phonological information and orthography information in the process of visual word recognition among the courses of language information processing, and the representation types of the mental lexicon. The model that this study is presenting here was designed as a feed forward network structure which is comprised of input layer which uses two Korean syllables as its input value, hidden layer, and output layer which express meanings. As the result of the study, the computational model showed the phonological and orthographic neighborhood effect among language phenomena which are shown in Korean word recognition, and showed proofs which implies that the mental lexicon is represented as phonological information in the process of Korean word recognition.

• Journal of Korean Library and Information Science Society
• /
• v.50 no.3
• /
• pp.97-116
• /
• 2019

This study is to analyze the characteristics of 2019 revised RDA and suggest the consideration in aspects of cataloging using the literature reviews. The following 3 things are suggested with analyzing the revised RDA. First, high quality data such as supplementing cataloging data and constructing vocabulary encoding schemes should be needed to transform bibliographic data to linked data for the semantic web. Second, MARC should be expanded to accept the new conept of LRM and linked data being reflected in revised RDA because MARC is the unique encoding format untile linked data will be transformed from MARC data. Third, the policy statement and the application profile are needed for describing resource consistently because each entity and element has own condition and option, and there are different elements for applying rules in revised RDA. Based on this study, the RDA related researches should be in progress such as exapanding BIBFRAME as well as MARC to accept the new concepts in revised RDA, and, also, reflecting and accepting RDA being able to use revised RDA rules and registries in libraries and nations that have been faced to revise their own cataloging rules.

• Annual Conference on Human and Language Technology
• /
• 2019.10a
• /
• pp.197-202
• /
• 2019

토큰화는 입력 텍스트를 더 작은 단위의 텍스트로 분절하는 과정으로 주로 기계 학습 과정의 효율화를 위해 수행되는 전처리 작업이다. 현재까지 자연어 처리 분야 과업에 적용하기 위해 다양한 토큰화 방법이 제안되어 왔으나, 주로 텍스트를 효율적으로 분절하는데 초점을 맞춘 연구만이 이루어져 왔을 뿐, 한국어 데이터를 대상으로 최신 기계 학습 기법을 적용하고자 할 때 적합한 토큰화 방법이 무엇일지 탐구 해보기 위한 연구는 거의 이루어지지 않았다. 본 논문에서는 한국어 데이터를 대상으로 최신 기계 학습 기법인 전이 학습 기반의 자연어 처리 방법론을 적용하는데 있어 가장 적합한 토큰화 방법이 무엇인지 알아보기 위한 탐구 연구를 진행했다. 실험을 위해서는 대표적인 전이 학습 모형이면서 가장 좋은 성능을 보이고 있는 모형인 BERT를 이용했으며, 최종 성능 비교를 위해 토큰화 방법에 따라 성능이 크게 좌우되는 과업 중 하나인 기계 독해 과업을 채택했다. 비교 실험을 위한 토큰화 방법으로는 통상적으로 사용되는 음절, 어절, 형태소 단위뿐만 아니라 최근 각광을 받고 있는 토큰화 방식인 Byte Pair Encoding (BPE)를 채택했으며, 이와 더불어 새로운 토큰화 방법인 형태소 분절 단위 위에 BPE를 적용하는 혼합 토큰화 방법을 제안 한 뒤 성능 비교를 실시했다. 실험 결과, 어휘집 축소 효과 및 언어 모델의 퍼플렉시티 관점에서는 음절 단위 토큰화가 우수한 성능을 보였으나, 토큰 자체의 의미 내포 능력이 중요한 기계 독해 과업의 경우 형태소 단위의 토큰화가 우수한 성능을 보임을 확인할 수 있었다. 또한, BPE 토큰화가 종합적으로 우수한 성능을 보이는 가운데, 본 연구에서 새로이 제안한 형태소 분절과 BPE를 동시에 이용하는 혼합 토큰화 방법이 가장 우수한 성능을 보임을 확인할 수 있었다.

• Journal of The Korean Association For Science Education
• /
• v.31 no.8
• /
• pp.1175-1185
• /
• 2011

Language skills in science education seem to have a great impact on student achievement. Most of the language skills affecting students' science learning can be accounted for an understanding of scientific terms. The ultimate goal of this study is selecting and grading the scientific terms for science education. As a basic research, we investigated scientific terms contained in science textbooks, because they are the most basic and selected terminologies. As a result of investigation of scientific terms in physics units of middle school science textbooks, we made a list of 556 scientific terms by grade level. This includes 249 words for grade 7, 170 words for grade 8, and 137 words for grade 9.