• Annual Conference on Human and Language Technology
• /
• 2009.10a
• /
• pp.44-49
• /
• 2009

통계기계번역에서 도메인에 특화된 번역을 시도하여 성능향상을 얻는 방법이 있다. 이를 위하여 문장의 유형이나 장르에 따라 클러스터링을 수행한다. 그러나 기존의 연구 중 문장의 유형 정보와 장르에 따른 정보를 동시에 사용한 경우는 없었다. 본 논문에서는 문장 사이의 문법적 구조 유사성으로 문장을 유형별로 분류하는 새로운 기법을 제시하였고, 단어 유사도 정보로 문서의 장르를 구분하여 기존의 두 기법을 통합하였다. 이렇게 분류된 말뭉치에서 추출한 모델과 전체 말뭉치에서 추출된 모델에서 보간법(interpolation)을 사용하여 통계기계번역의 성능을 향상하였다. 문장구조의 유사성과 단어 유사도 계산을 위하여 각각 커널과 코사인 유사도를 적용하였으며, 두 유사도를 적용하여 말뭉치를 분류하는 과정은 K-Means 알고리즘과 유사한 기계학습 기법을 사용하였다. 이를 일본어-영어의 특허문서에서 실험한 결과 최선의 경우 약 2.5%의 상대적인 성능 향상을 얻었다.

• Annual Conference on Human and Language Technology
• /
• 2015.10a
• /
• pp.23-28
• /
• 2015

본 논문에서는 한국어 트리플 생성 시스템의 정확도를 향상시키기 위한 distant supervision 기반의 신뢰도 측정 방법을 제안한다. 기존의 많은 패턴 기반의 트리플 생성 시스템에는 distant supervision의 기본 가정으로 인해 다수의 오류 패턴이 발생할 여지가 크다. 기존의 연구에서는 오류 패턴을 제거하기 위하여 발생 빈도, 공기 횟수 등의 통계에 기반하여 간접적으로 신뢰도를 측정하였다. 본 논문에서는 한국어 패턴과 영어 프로퍼티 사이의 의미 유사도를 측정함으로써 통계에 기반한 방법보다 더 정확한 신뢰도 측정 방법을 제안한다. 비지도 학습 방법인 워드임베딩을 활용하여 어휘의 의미를 학습하고, 이들 사이의 유사도를 측정한다. 한국어 패턴과 영어 프로퍼티의 어휘 불일치 문제를 해결하기 위하여 정준상관분석을 활용하였다. 실험 결과에 따르면 본 논문에서 제안한 패턴 신뢰도 측정 방법은 통계 기반의 방법에 비해 정확률이 9%나 더 높은 트리플 집합을 생성함을 보여주어, 의미 유사도를 반영한 신뢰도 측정이 기존의 통계 기반 신뢰도 측정보다 고품질 트리플 생성에 더 적합함을 확인하였다.

• Journal of the Korea Organic Resources Recycling Association
• /
• v.29 no.4
• /
• pp.15-24
• /
• 2021

This study was conducted to evaluate effects of microbial fertilizer (MF) containing Bacillus amyloliquefaciens on the growth in the lettuce by treating MF without and with organic fertilizer (OF), or by its formulation types, and to investigate its application in the eco-friendly agriculture. B. amyloliquefaciens, active microbe of MF, had activities of amylase and protease. Applied only MF without OF, MF treatments were not significantly different with non-fertilizer (NF). As compared to control, dry weight of MOF2 treatment (2,500 kg OF/ha + 50 kg MF/ha) was increased by about 30%. As applied with wettable powder type (WP) and soluble powder type (SP) of MF, the dry weight of WP was increased by 43% than that of control, but SP not significantly different. In the comparison with two MF formulation, dry weight of WP was increased by about 37% than that of SP. These results indicated that an application of MF improved the growth of lettuce by prompting a mineralization of OF, and that the formulation type of MF was better WP than SP.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.17 no.7
• /
• pp.719-729
• /
• 1992

An improved MXTM (maximum trimmed mean) -CFAR (constant false alarm rate) processor is proposed to reduce false alarm rates In detecting radar targets and Its performance character is ticsare analyzed to be compared with those of other CFAR processors. The proposed MXTM-CFAR processor is obtained by combining the GO (greatest of ) -CFAR processor reducing excessive falsealarm rate at riutter edges with the TM-CFAR processor showing good performances In homo-geneous Jnonhornog eneous background. Performance analyses have been done by computing detection probability, constant false alarm rate and detection thresholds under the homogeneous or multiple target environments and at the clutter edges. Analysis results how that the proposed CFAR processor maintains its performance as good as those of,05(order statistics) and TM-CFAR inhomogeneous and multiple target environments and Can reduce the false alarm rate at clutter edges. Overall computing time hfs been also reduced.

• Journal of Korean Society of Disaster and Security
• /
• v.15 no.3
• /
• pp.79-88
• /
• 2022

Cable-supported bridges, as important large infrastructures, require a long-term and systematic maintenance strategy. In particular, various methods have been proposed to secure safety for the bridges, such as installing various types of sensor on members in the bridges, and setting management thresholds. It is evidently necessary to propose a strategic plan to efficiently manage increasing number of cable-supported bridges and data collected from a number of sensors. This study aims to develop an analysis tool that can automatically remove abnormal signals and calculate statistical results for the purpose of efficiently analyzing a wide range of data collected from a long span bridge measurement system. To develop the tool, basic information such as the types and quantity of sensors installed in long span bridges and signal characteristics of the collected data were analyzed. Thereafter, the Humpel filtering method was used to determine the presence or absence of an abnormality in the signal and then filtered. The statistical results with filtered data were shown. Finally, one cable-stayed bridge and one suspension bridge currently in use were chosen as the target bridges to verify the performance of the developed tool. Signal processing and statistical analysis with the tool were performed. The results are similar to the results reported in the existing work.

• Journal of Food Hygiene and Safety
• /
• v.35 no.3
• /
• pp.284-289
• /
• 2020

This study was performed with hot water extract (LJF-W) and 70% ethanol extract (LJF-70E) of Litsea japonica fruit to investigate the immunostimulatory activity and bone health-promoting effect of L. japonica fruit. The production of pro-inflammatory mediator (nitric oxide) and pro-inflammatory cytokines (TNF-α, IL-6 and IL-1β) in murine macrophage RAW264.7 cells were estimated to examine the immunostimulatory activity of the fruit extracts. The immunostimulatory activity of LJF-W was higher than that of positive control (geinsenosides). However, there was no effect in LJF-70E. Futhermore, both LJF-W and LJF-70E appeared to stimulate the proliferation of MC3T3-E1 cells, proving the effect as a bone health agent. From this result it could be presumed that L. japonica fruit extracts not only stimulate the immune system, but also the ability to promote bone health.

• The Journal of Korean Academy of Prosthodontics
• /
• v.56 no.3
• /
• pp.199-205
• /
• 2018

Purpose: This study evaluated the effect of nonthermal plasma treatment on the bond strength of autopolymerizing relining resin to the injection molded thermoplastic denture base resins (TDBRs) with different surface treatments. Materials and methods: Acrylic Resin (Acrytone), Polyester (Estheshot-Bright), Polyamide (Valplast) and Polypropylene (Weldenz) were subjected to various surface treatments: No treatment, Nonthermal plasma, Sandblasting, Sandblasting and nonthermal plasma. Specimens were bonded using an autopolymerizing relining resin. Shear bond strength was tested using universal testing machine with crosshead speed of 1 mm/min. Statistical analysis by two-way analysis of variance with Tukey's test post hoc was used. Results: Acrytone showed significantly higher shear bond strength value among other TDBR group while Weldenz had the lowest. The sandblasting and nonthermal plasma condition had significantly higher shear bond strength value in all of the resin groups (P < .05). Conclusion: The use of nonthermal plasma treatment showed limited effect on the shear bond strength between TDBRs and relining resin, and combination of nonthermal plasma and sandblasting improved the shear bond strength between TDBR and reline material.

• Journal of dental hygiene science
• /
• v.2 no.2
• /
• pp.105-113
• /
• 2002

This study was aimed to investigate the practice on the infection control and dental waste disposal in dental clinic. The data used in this study had been collected in 100 dental clinics located in Seoul city from August to October, 2001. It was shown that examination instruments, oral surgery instruments, and periodontal instruments were almost sterilization in dental clinic. It was high in the recognition of individual protection but it was low in the degree of practice. Therefore, the practice in infection control is very important to dental team for safety of patients and operators.

• Korean Journal of Poultry Science
• /
• v.35 no.4
• /
• pp.399-405
• /
• 2009

Two experiments were conducted to establish the basic data of feeding system in cross-bred Korean native chicks. A total of seven hundred twenty chicks were replaced the 36 floor pen for the first five weeks with $3{\times}3$ factorial design in Experiment 1. Four hundred eighty of five weeks old chicks were raised from six to ten weeks of age with $3{\times}2$ factorial design in Experiment 2. Dietary ME and CP were formulated to contain 3,000, 3,050, 3,100 kcal/kg and 21, 22, 23%, respectively in Experiment 1 and 3,050, 3,100, 3,150 kcal/kg and 18, 19% in Experiment 2. Weight gain, feed intake were measured and calculated the feed conversion. Blood were collected and analyzed at the end of experiments. In Experiment 1, weight gain showed significantly higher in 3,050, 3,100 kcal/kg treatments than 3,000 kcal/kg treatment (P<0.05), but was not different in CP treatments. Feed intake was statistically high in 3,000 kcal/kg treatment compared with 3,050 and 3,100 kcal/kg ones (P<0.05), and more increased in 21% CP treatments compared to that of 22 and 23 CP treatment (P<0.05). Feed conversion of birds fed 3,050 and 3,100 kcal/kg diet showed much lower than 3,000 kcal/kg treatments (P<0.05). FCR was signicantly improved (P<0.05) in chicks fed diets containing 21 and 22% CP as compared to that fed 20% CP. Blood protein, glucose, and total cholesterol tended to increase in high energy and diet treatments. Blood HDL was increased as dietary energy increased, whereas LDL increased in low CP treatments. In Experiment 2, weight gain was not consistent between treatments, but more increased in 18% CP treatments compared to that of 19% CP treatment from six to ten weeks old in cross bred chicks (P<0.05). Feed intake was similar to the result of weight gain, but more increased in 19% CP treatment than 18% CP treatment (P<0.05). There were no statistically difference in FCR, but seemed to improve as dietary ME increased. Blood total protein and glucose increased as dietary CP was high, but triglyceride and HDL increased in high versus low ME (P<0.05). The results of these experiments suggested that optimum dietary ME and CP, were 3,050, 3,150kcal/kg and 22, 19% for the first five weeks and second one, respectively.

• The Transactions of the Korea Information Processing Society
• /
• v.4 no.4
• /
• pp.1035-1045
• /
• 1997

It is important to understand how we can deal with doements for the modeling of neural networks when we are unbestigating the dynamical performance and the information procoessing capabilitids.The information procewssing capabkities of model neural networks will change for different response, synaptic weights or learning rules. Using the staritical neurodyamics method, we evalute the capabikities of neural networks in order to understand the basic conept ofr parallel distributed processing. In this paper, we explain the reuslts of theoretical anaysis of the effests of the hysteretic property on the performance of wuquential associative neral networks.