• Proceedings of the Korean Statistical Society Conference
• /
• 2000.11a
• /
• pp.141-146
• /
• 2000

Shapiro와 Wilk(1972)는 위치모수와 척도모수가 미지인 경우 지수분포의 검정통계량을 제안하였다. 그것은 척도모수의 일반화 최소제곱추정량과 표본분산의 비로 구성되었다. 그러나 이 검정통계량은 일치성을 갖지 않는다. 본 논문에서는 척도모수의 두개의 점근유효추정량으로 구성된 통계량을 고려하고 이의 극한분포를 구하였다. 또한 두 개의 통계량의 검정력을 비교한 결과 제안된 통계량이 변동계수가 1보다 크거나 같은 분포에서 더 좋은 검정력을 가짐을 볼 수 있었다.

• Korean Journal of Heritage: History & Science
• /
• v.52 no.1
• /
• pp.44-63
• /
• 2019

This study empirically investigated the construction and aspects of change in Chugyeongwon, which is located in Donggweol (東闕). In detail, this study investigated the location of the construction and range of Chugyeongwon, the background and intention of the construction, and the affiliated system and aspects of spatial changes of it. The research results can be summarized as follows: First, Chugyeongwon has been assumed to be the space near Haminjeong (涵仁亭) or between Simindang (時敏堂) and Jinsudang (進修堂) in Changgyeonggung Palace. However, according to related historical materials, it is said that Chugyeongwon was located west of Dochongbu (都摠府) in Hyeopsangmun (協祥門) and near Sungmundang (崇文堂). Through Donggweoldohyeong (東闕圖形), evidence of the construction of Chugyeongwon can be found, which verifies such claims. According to The Plan of Changgyeonggung Palace (昌慶宮配置圖), in the form of modern measured drawing, Chugyeongwon today is the green space created in the south of Munjeongjeon (文政殿) and Sungmundang in Changgyeonggung Palace. Second, According to Donggweoldo (東闕圖), Chugyeongwon was a green space where trees grew on the ground within the walls. No artificial facilities were constructed inside. In addition, Chugyeongwon was located at a site with an altitude higher than the surroundings. Especially, the composition forms and location characteristics of Chugyeongwon are similar to those of the Palace Outer Garden located in Hanyang. Thus, based on this evidence about the form and other aspects of the operation of the Palace Outer Garden, it can be inferred that Chugyeongwon was constructed for the preservation and cultivation of the geographical features inside Donggweol. Third, in the late Joseon period, Chugyeongwon was assigned to Changdeokgung Palace or Changgyeonggung Palace in the same manner as was Donggung (東宮). Thus, it is very likely that Chugyeongwon served as a garden for the Royal Family in the Donggung area. The west boundary of Chugyeongwon, which originally consisted of walls and a side gate, was changed into the form in which the walls and colonnades were combined. Chugyeongwon has been modified due to various acts of development since the Japanese colonial era, and in the end, it has disappeared so that no trace can be found.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2016.05a
• /
• pp.106-107
• /
• 2016

Rolling motion of floating body might upset the body, make crews and passengers exhausted and/or apply forces to the structure to cause damage. Therefore for almost ships bilge keels are equipped, in special case fin stabilizer or gyroscope may be installed. This paper suggests the Anti-rolling pendulum to reduce roll motion to act the similar role with anti-rolling tank. The device suggested has more effective than the anti-rolling tank with 1/6 volume of the tank.

• Survey Research
• /
• v.5 no.2
• /
• pp.49-70
• /
• 2004

This paper is a broad review about general regression estimators, which are very useful when auxiliary variables are available in survey sampling. We investigate the process of development of general regression estimators from birth to suggestion of variance estimation method and examine some properties of general regression estimators by comparing with calibration and QR estimators. We also present some forms of general regression estimators available under complex sampling designs such as stratified sampling and cluster sampling. Finally, we comment some advantages as well as disadvantages of general regression estimators and theoretical and practical development in the future.

• Annual Conference on Human and Language Technology
• /
• 2019.10a
• /
• pp.415-418
• /
• 2019

기계독해 모델에 새로운 도메인을 적용하기 위해서는 도메인에 맞는 데이터가 필요하다. 그러나 추가 데이터 구축은 많은 비용이 발생한다. 사람이 직접 구축한 데이터 없이 적용하기 위해서는 자동 추가 데이터 확보, 도메인 적응의 문제를 해결해야한다. 추가 데이터 확보의 경우 번역, 질의 생성의 방법으로 연구가 진행되었다. 그러나 도메인 적응을 위해서는 새로운 정답 유형에 대한 질의가 필요하며 이를 위해서는 정답 후보 추출, 추출된 정답 후보로 질의를 생성해야한다. 본 논문에서는 이러한 문제를 해결하기 위해 듀얼 포인터 네트워크 기반 정답 후보 추출 모델로 정답 후보를 추출하고, 포인터 제너레이터 기반 질의 생성 모델로 새로운 데이터를 생성하는 방법을 제안한다. 실험 결과 추가 데이터 확보의 경우 KorQuAD, 경제, 금융 도메인의 데이터에서 모두 성능 향상을 보였으며, 도메인 적응 실험에서도 새로운 도메인의 문맥만을 이용해 데이터를 생성했을 때 기존 도메인과 다른 도메인에서 모두 기계독해 성능 향상을 보였다.

• The Journal of the Acoustical Society of Korea
• /
• v.29 no.7
• /
• pp.422-430
• /
• 2010

The radiation property of a Tonpilz transducer is influenced by the structural variables of the transducer. In this study, with respect to a single mode Tonpilz transducer, the radiation patterns were calculated for different head mass geometries of the same effective radiation area. The shapes of the head mass analyzed were the most popular circular, regular triangular, square, regular hexagonal and regular octagonal types, and radiation pattern equations were derived for each of the head mass shapes. Based on the derived equations, radiation patterns in accordance with the shape and size of the head mass were calculated and the results were compared with each other. Validity of the calculation results were confirmed by means of finite element analysis.

• Journal for History of Mathematics
• /
• v.19 no.3
• /
• pp.113-122
• /
• 2006

We investigate the mathematical background, statistical methodology, and the teaching of computer-software field using smoothing methodology in this paper. Also we investigate conception and methodology of histogram, kernel density estimator, adaptive kernel estimator, bandwidth selection based on mathematics and statistics.

• Annual Conference on Human and Language Technology
• /
• 2021.10a
• /
• pp.396-401
• /
• 2021

초거대 언어모델를 활용한 퓨샷(few shot) 학습법은 여러 자연어 처리 문제에서 좋은 성능을 보였다. 하지만 데이터를 활용한 추가 학습으로 문제를 추론하는 것이 아니라, 이산적인 공간에서 퓨샷 구성을 통해 문제를 정의하는 방식은 성능 향상에 한계가 존재한다. 이를 해결하기 위해 초거대 언어모델의 모수 전체가 아닌 일부를 추가 학습하거나 다른 신경망을 덧붙여 연속적인 공간에서 추론하는 P-tuning과 같은 데이터 기반 추가 학습 방법들이 등장하였다. 본 논문에서는 문맥에 따른 질의 정규화 문제를 대화형 음성 검색 서비스에 맞게 직접 정의하였고, 초거대 언어모델을 P-tuning으로 추가 학습한 경우 퓨샷 학습법 대비 정확도가 상승함을 보였다.

• Journal of the Korean Society of Fisheries and Ocean Technology
• /
• v.38 no.1
• /
• pp.20-35
• /
• 2002

We studied the fishing ground of a yellowtail and oceanographic conditions around Chujado and Marado around Jeju Island from autumn to winter. Also we investigate the forming mechanism of fishing ground around Marado, which shows the best catch of yellowtail near Jeju island. The obtained results are summarized as follow ; 1. When the high temperature water about 15.0∼19.0 in depth of 50m distributed around Chujado and Marado, the fishing ground of Yellowtail is shows a good catch. On the contrary, when the low temperature water about 11.0∼14.0 is shows a poor catch. 2. Fishing grounds of yellowtail by handline fishing are formed with a school of Yellowtail migrating southward from the coastal sea of Korea and slaying around Marado to winter and spawn due to following factors. Firstly, a front of temperature and salinity is formed between inshore water and oceanic water around Marado. Secondly, small-size eddies and excellent horizontal and vortical mix which is created by strong tide and geographical features including irregular underwater ground and an isolated island. An abundant school of yellow tails stays in front of the tide where a good fishing ground was made.

• Communications for Statistical Applications and Methods
• /
• v.5 no.3
• /
• pp.623-632
• /
• 1998

이원혼합모형에서 고정효과의 추정가능한 함수에 대한 신뢰구간을 구하는 경우에 어떤 분산성분추정량을 선택하는 것이 가장 바람직한가를 모의실험을 통하여 살펴본다 혼합모형에서는 t-분포와 일반화최소제곱추정량을 사용하여 신뢰구간을 구할 수 있는데, 일반적으로 분산성분을 알 수 없기 때문에 분산성분을 반드시 추정하여야만 한다. 이 경우 분산성분의 추정량으로 가장 많이 사용되는 추정량들인 Henderson의 방법 III 추정량, 사전추측값이 1인 MINQUE 추정량, MLE(최우추정량), REMLE(제한최우추정량)를 이용하여 분산행렬을 추정하고, 신뢰구간의 포함범위확률과 평균길이를 모의실험을 통하여 살펴본다. 모의실험의 결과는 4가지 추정량 모두 비슷한 신뢰구간의 포함범위확률과 평균길이를 갖는 것으로 판명되었다.