• Annual Conference on Human and Language Technology
• /
• 2019.10a
• /
• pp.246-251
• /
• 2019

품사 부착 말뭉치란 문장에 포함된 각 단어에 품사 표지를 부착한 말뭉치를 말한다. 이런 말뭉치에는 다양한 형태의 오류들이 포함되어 있으며, 오류가 포함된 말뭉치를 학습 자료로 사용하는 자연언어처리 시스템의 좋은 성능을 기대할 수 없다. 따라서 말뭉치의 일관성이나 정확도는 자연언어처리 시스템의 성능에 많은 영향을 준다. 하지만 말뭉치 구축 과정에서 작업자의 실수가 발생하고 여러 작업자가 작업을 수행하다 보니 일관성을 유지하기가 쉽지 않다. 본 논문에서는 이러한 문제를 해결하기 위해서 GMM을 이용한 군집화를 수행하여 오류 후보를 추출한다. 이를 통해서 말뭉치 구축 과정에서 작업자의 실수를 방지하고 일관성을 유지하고자 한다. 세종품사부착 말뭉치를 대상으로 임의로 오류를 유발시켜 실험한 결과, 재현율 84.74%의 성능으로 오류를 탐지하였다. 향후에 좀 더 높은 재현율을 위해서 자질 확장이나 회귀 분석 방법 등을 추진할 계획이다.

• 재활복지
• /
• v.20 no.3
• /
• pp.67-82
• /
• 2016

The purpose of this study was to investigate relation and effect of PCC(Percentage of Correct Consonant) and nasalance score on the speech intelligibility and acceptability in adults with dysarthria by reading task of standardized passage. Ten adults with dysarthria and sixteen normal adults were participated in this study. PCC and nasalance score were measured through reading task of standardized passage. And, speech intelligibility and acceptability were examined using visual analogue criteria. The result of the study was as follows. First, the nasalance score of adults with dysarthria group is significantly higher than normal adults group in reading sample by standardized passage. Second, the PCC, speech intelligibility and acceptability shows significant correlation. However, the nasalance score doesn't show significant corelation with speech intelligibility and acceptability. These results suggest that PCC is closely related to speech intelligibility and speech acceptability, but nasalance score is not related to speech intelligibility and speech acceptability.

• Journal of Korean Classical Literature and Education
• /
• no.32
• /
• pp.29-61
• /
• 2016

A joke (才談, 재담) is "the most interesting and witty language unit" in our speech. However, the search of a joke is still starting. Although joke are related to the witty and interesting talks, stories, songs and plays, the actual object of a joke is only the witty and interesting talk. A joke is witty talk that is interesting or laughter-inducing. Many Jokes can be found in the traditional Korean funny performing arts (演戱, 연희). This is because these art forms are performed in open yards, which necessitated amusing the audience, amusement, in its turn, required jokes. Jokes in the traditional funny performing arts can generally be classified as follows: 1) Jokes related to a situation: These include right words at a given situation, exaggerating words, diminishing words, deviancy words, and cause-effect words. 2) Jokes related to discourse: These include enumerating words, amplificatory words, contrasting words, fluently lying words, undeniable words, purposely unknowing words, and deliberately incorrect words. 3) Jokes related to vocabulary: These include synonym, similar words, changed word-ordering words, and incorrect words. 4) Jokes related to pronunciation: These include homonyms, and anti-homonyms. Although there may be other jokes, those presented above are typical ones. A joke is "the result that human being can achieve when he/she has overcome natural and social difficulties and is left with only a free and creative spirit." Jokes are necessary in all ages and everywhere. Today, more varied and high-level jokes can be created by developing the diversity of jokes in traditional funny performing arts. Also, I expect new sorts of jokes, because a joke always demands a creative spirit.

• Journal of the Society of Cosmetic Scientists of Korea
• /
• v.46 no.1
• /
• pp.31-42
• /
• 2020

Horse fat is known to be an effective ingredient in Asia, and the horse fat itself, which is mixed with other ingredients at the additive level, is often sold as a finished product. In this case, physical properties of the horse fat raw material are important. Many horse fats produced in Korea (Jeju) have low temperature stability, so if not stored at low temperatures, segregation may occur. In the case of Japanese horse fat, it is partially hydrogenated or is used the solid phase as the horse fat by separating the liquid phase and the solid phase that is harder and more stable than the horse fat of Jeju. In this study, the physical properties were tested to improve the temperature stability even without the partial hydrogenation process of Jeju horse fat. Various oil gelling agents were used in the study. Results confirmed that the physical properties of the hydroxystearic acid added Jeju horse fat were improved. In addition, stability evaluations at temperatures of 25 ℃, 40 ℃, 45 ℃ and flow behavior evaluations at temperatures of 25 ℃, 30 ℃, 40 ℃ were performed for Jeju horse fat with hydroxystearic acid, 100% Jeju horse fat, and 100% Japanese horse fat. Results showed that the Jeju horse fat improved in flow behavior by adding hydroxystearic acid similar to that of Japanese horse fat. In addition, when the crystal state was observed under a microscope, the thermal stability was improved by decreasing the size of the needle-type crystals with the addition of hydroxystearic acid. Jeju horse fat containing hydroxystearic acid was found to have no physical problems even when stored at room temperature for a long time.

• Journal of the korean veterinary medical association
• /
• v.34 no.2
• /
• pp.124-130
• /
• 1998

For every person who concerned with horse, horse breeder, groom and all the horse related vet, oedema, on horse's articulation, ligament, tendon, lacertus or ligamenta annularia, is unavoidable disease. It is the horse in obest condition, so called 'clear

• Journal of KIISE
• /
• v.43 no.6
• /
• pp.636-644
• /
• 2016

Annotated Corpus is important to understand natural language using machine learning method. In this paper, we propose a new method to automate error reduction of annotated corpora. We use the Ripple-Down Rules(RDR) for reducing errors and Kernel to extend RDR for NLP. We applied our system to the Korean Wikipedia and blog corpus errors to find the annotated corpora error type. Experimental results with various views from the Korean Wikipedia and blog are reported to evaluate the effectiveness and efficiency of our proposed approach. The proposed approach can be used to reduce errors of large corpora.

• Journal of KIISE:Computing Practices and Letters
• /
• v.16 no.11
• /
• pp.1046-1050
• /
• 2010

Corpus is widely used as a fundamental resource for various purposes in linguistic studies. There are several large corpora such as Sejong corpus in Korea. However, it is hard to find a tool utilizing such large corpora. In this paper, we propose a method of utilizing Sejong corpus based on the relational database. We designed the relational database scheme to store corpus and implemented a Web-based application so that many researchers can easily access and utilize the Sejong corpus.

• Communications of Mathematical Education
• /
• v.30 no.2
• /
• pp.121-138
• /
• 2016

Most of elementary and secondary school mathematical terms in Korean are Sino-Korean words. We check Chinese characters relating to such Sino-Korean words by using Chinese dictionaries, and critically judge how much we can understand Sino-Korean words by Chinese characters. Through this search, we classify Sino-Korean words into three categories; words which can be understood by Chinese characters, words which can not be understood by Chinese characters, words which are misunderstood by Chinese characters.

• Proceedings of the Korean Information Science Society Conference
• /
• 2003.04c
• /
• pp.483-485
• /
• 2003

구문분석 말뭉치는 통계적 구문분석 분야의 필수적인 항목으로 많은 유용성을 가지지만, 말뭉치를 구축할 때 막대한 시간과 비용이 요구되기 때문에 구축자의 수작업을 감소시키는 방법에 대한 연구가 필요하다. 본 논문에서는 대량의 신뢰도 있는 구문분석 말뭉치를 구축하기 위해 신경망을 사용하는 반자동 구문 분석 말뭉치 구축도구에 대해서 설명한다. 개발된 도구는 구문패턴 추골, 신경망 학습, 반자동 구축의 세 단계로 구성된다. 구문패턴 추출 단계에서는 사용자가 정의한 자질집합을 사용하여 기존에 구축된 말뭉치에서 구문패턴들을 추출하고, 신경망 학습의 단계에서는 추출된 구문패턴들을 사용하여 신경망을 학습한다. 그리고, 반자동 구축 단계에서는 학습된 신경망을 사용하여 반자동으로 구문분석 말뭉치를 구축한다. 본 논문에서 제안하는 방법은 다양한 자질집합을 조합하여 사용할 수 있고, 학습을 사용하기 때문에 학습 집합에 나타나지 않은 경우에 대해서도 합리적인 결정을 내릴 수 있다. 소량의 구문분석 말뭉치를 대상으로 실험한 결과, 본 논문에서 제안하는 방법이 약 42.5%의 수작업 횟수 감소율을 보였음을 알 수 있었다.

• 보건세계
• /
• v.43 no.12 s.484
• /
• pp.4-9
• /
• 1996

기도가 협착되어서 증상을 일으키는 질환으로 만성기관지염, 폐기종, 만성폐쇄성질환, 기관지천식 등이 있으며, 기침, 가래, 호흡곤란(숨참)등이 주요 증상인데 비의료인에게는 해소천식으로 더 잘 알려져 있다. 일반적으로 기침 가래가 나오고 숨이차면 해소천식이라고 하는데 실제로 이 만큼 오용되고 남용되는 단어는 드물 것이다. 동의보감에 의하면 해수(해소의 원명)는 기침을 말하는데 해는 가래가 없이 소리가 심한 경우이고 수는 가래를 주로하는 기침을 말하며 16가지 종류로 구분된다. 천식은 숨이 찬 증상을 말하는 것으로 8가지 종류로 구분되어 있다. 따라서 해소천식은 기침가래가 나오고 숨이 찬 증상을 말하는 것으로 특정질환을 말하는 병명이 아니다. 비의료인이 말하는 천식의 경우에는 숨이 찬 증상을 말하는 것이고 의료인이 천식이라고 할 경우에는 특정 기도질환인 기관지 천식을 말하는 것이므로 서로 오해하기가 쉽다. 실제로 일간지 건강상식란에 천식을 소개하였는데 의사는 병명으로서의 기관지천식을 설명하였고 동참한 한의사는 증상으로서의 천식을 설명하여 혼란을 일으켰던 예가 있었다. 한의학에서는 예전에 기관지천식은 효천(孝喘)이라고 부르고 천식의 증상이 심하고 위중한 상황일 때에는 천급(喘急)이라고 하여 한의학에서도 구별하였는데 도리어 현대에 와서 정확성이 없어지고 모호해진 이유는 무엇일지 궁금하다. 폐쇄성기도질환은 기도가 좁아지는 기전과 기도폐쇄의 가역성(기도가 정상으로 회복되는지)여부에 다라서 가역성폐쇄질환인 기관지천식과 비가역적인 만성폐쇄성기도질환으로 구분하고 있다.