• 한국환경복원기술학회지
• /
• 제14권2호
• /
• pp.47-57
• /
• 2011

This study monitored forest plant species vulnerable to climate change in Jiri Mountain, one of Korea's representative alpine regions, in order to securely preserve plant genetic resources susceptible to climate change and to utilize the results as basic data for bioclimatology prediction and management on a long-term basis. A majority of indicator plants tended to blossom one week to one month later in 2010 than in 2009. As with the blooming dates, the falling dates of blossoms became later in most species, with the exception for Weigela florida and Oplopanax elatus. Leaf bursting as well fell on later dates in a majority of species excluding Carpinus laxiflora and Cupressus sempervirens, displaying the most obvious differences among the data of analysis of the 2009-2010 physiological cycle changes. It is believed that was due to the fact that temperatures in February, March and April, which affect plants' blossoming and leaf bursting, were lower in 2010 than in 2009 and that cold temperatures in the winter lasted for a longer period in 2010 than in 2009. The dates of leaves being changed to red were similar in 2009 and 2010 by being or later or earlier by several weeks in 2010 than in 2009 without any regularity. Most species' leaves began to fall at similar dates in 2009 and 2010 or at later dates by one to two weeks in 2010 than in 2009. The temperature differences in late 2009 and late 2010 were not so large, resulting in similar dates of falling leaves, and gaps in several indicator plants' physiological cycles without any regularity can be attributed to each individual plant's physiological and environmental characteristics.

• 한국초지조사료학회지
• /
• 제37권4호
• /
• pp.350-357
• /
• 2017

Near infrared spectroscopy (NIRS) is a rapid and accurate method for analyzing the quality of cereals, and dried animal forage. However, one limitation of this method is its inability to measure fermentation parameters in dried and ground samples because they are volatile, and therefore, respectively lost during the drying process. In order to overcome this limitation, in this study, fresh coarse haylage was used to test the potential of NIRS to accurately determine chemical composition and fermentation parameters. Fresh coarse Italian ryegrass haylage samples were scanned at 1 nm intervals over a wavelength range of 680 to 2500 nm, and optical data were recorded as log 1/reflectance. Spectral data, together with first- and second-order derivatives, were analyzed using partial least squares (PLS) multivariate regressions; scatter correction procedures (standard normal variate and detrend) were used in order to reduce the effect of extraneous noise. Optimum calibrations were selected based on their low standard error of cross validation (SECV) values. Further, ratio of performance deviation, obtained by dividing the standard deviation of reference values by SECV values, was used to evaluate the reliability of predictive models. Our results showed that the NIRS method can predict chemical constituents accurately (correlation coefficient of cross validation, $R_{cv}^2$, ranged from 0.76 to 0.97); the exception to this result was crude ash ($R_{cv}^2=0.49$ and RPD = 2.09). Comparison of mathematical treatments for raw spectra showed that second-order derivatives yielded better predictions than first-order derivatives. The best mathematical treatment for DM, ADF, and NDF, respectively was 2, 16, 16, whereas the best mathematical treatment for CP and crude ash, respectively was 2, 8, 8. The calibration models for fermentation parameters had low predictive accuracy for acetic, propionic, and butyric acids (RPD < 2.5). However, pH, and lactic and total acids were predicted with considerable accuracy ($R_{cv}^2$ 0.73 to 0.78; RPD values exceeded 2.5), and the best mathematical treatment for them was 1, 8, 8. Our findings show that, when fresh haylage is used, NIRS-based calibrations are reliable for the prediction of haylage characteristics, and therefore useful for the assessment of the forage quality.

• 한국전문물리치료학회지
• /
• 제12권3호
• /
• pp.74-83
• /
• 2005

The purpose of this study was to produce the regression equation from non-exercise $VO_{2max}$ of healthy young adults and to develop a maximal oxygen consumption ($VO_{2max}$) regression model. This model was based on heart rate non-exercise predictor variables (rest heart rate, maximal heart rate/rest heart rate), as an extra addition to the general regression which can reflect an individual's inherent or acquired cardiorespiratory fitness. The subjects were 101 healthy young adults aged 19 to 35 years. Exercise testing was measured by using a Balke protocol for treadmill and indirect calorimetry. The prediction equation was analyzed by using stepwise multiple regression procedures. The mean of $VO_{2max}$ was $39.02{\pm}6.72\;m{\ell}/kg/min$ (mean${\pm}$SD). The greatest variable correlated to $VO_{2max}$ was %fat. The predictor variable used in the non-exercise $VO_{2max}$ included %fat, gender, habitual physical activity and $HR_{max}/HR_{rest}$. The non-exercise $VO_{2max}$ estimation was as follows: $VO_{2max}$($m{\ell}/kg/min$)=55.58-.41(%fat)+.59(physical activity rating)-2.69($HR_{max}/HR_{rest}$)-5.36 (male=0, female=1); (R=.85, SEE=3.64, R2=.72: including heart rate variable); $VO_{2max}$($m{\ell}/kg/min$)=48.47-.41(%fat)+.45(physical activity rating)-5.12 (male=0, female=1); (R=.84, SEE=3.74, R2=.70: with the exception of heart rate variable). As an added heart rate variable, there was only a 2% coefficient of determination improved. Therefore, these results demonstrated that heart rate variable correlation with a non-exercise regression model was very low. In conclusion, for healthy young korean adults, those variables that can affect non-exercise $VO_{2max}$ estimation turned out to be only % fat, gender, and physical activity. We suggest that further research of predictor variables for non-exercise $VO_{2max}$ is necessary for different patient groups who cannot perform maximal exercise or submaximal exercise.

• 한국도로학회논문집
• /
• 제18권4호
• /
• pp.19-28
• /
• 2016

PURPOSES : Long-life asphalt pavements are used widely in developed countries. In order to be able to devise an effective maintenance strategy for such pavements, in this study, we evaluated the performance of the long-life asphalt pavements constructed along the national highways in South Korea. Further, an economic evaluation of the long-life asphalt pavements was performed based on a life-cycle cost analysis. We aimed to devise a model for evaluating the performance of long-life asphalt pavements using the national highway pavement management system (PMS) database as well as for analyzing the economic feasibility of such pavements, in order to promote their use in South Korea. METHODS : The maintenance history and pavement performance data were obtained from the national highway PMS database. The pavement performances for a total of 292 sections of 10 lanes (5 northbound lanes and 5 eastbound lanes) of national highways were used in this study. Models to predict the performances of hot mix asphalt (HMA) and long-life asphalt pavements under two distinct traffic conditions were developed using a simple regression method. Further, the economic feasibility of long-life asphalt pavements was evaluated using the Korea Pavement Management System (KoPMS). RESULTS : We developed service-life prediction models based on the traffic volume and the equivalent of single-axle load and found that long-life asphalt pavements have service lives 50% longer than those of HMA pavements. Further, the results of the economic analysis showed that long-life asphalt pavements are superior in terms of various economic indexes, including user cost, delay cost, total cost, and user benefits, even though their maintenance cost is higher than that of HMA pavements. A comparison of the economic feasibilities of the various groups showed that group A is superior to HMA pavements in all aspects except in terms of the maintenance criterion (crack 20% or higher) as per the NPV index. However, the long-life asphalt pavements in group B were superior in terms of the maintenance criterion (crack 25% or higher) regardless of the economic feasibility. CONCLUSIONS : The service life of long-life asphalt pavements was found to be approximately 50% longer than that of HMA pavements, regardless of the traffic volume characteristics. The economic feasibility of long-life asphalt pavements was evaluated based on the KoPMS. The results of the economic analysis were the following: long-life asphalt pavements are exceptional in terms of almost all factors, such as user cost, delay cost, total cost, and user benefit; however, the exception is the maintenance cost. Further, the economic feasibility of the long-life asphalt pavements in group B was found to be better than that of the HMA pavements (crack 25% or higher).

• 한국토양비료학회지
• /
• 제23권3호
• /
• pp.167-172
• /
• 1990

우리나라 토양(土壤)(제주도(濟州道) 토양(土壤) 제외(除外))의 토성별(土性別) 유효수분함량(有效水分含量) 평균치(平均値)를 구(求)하고 토양성질(土壤性質) 사이의 상관(相關)을 구명(究明)하고 유효수분함량(有效水分含量)의 추정식(推定式)을 얻기 위하여 시(市) 군(郡) 대표단면(代表斷面)의 토양분석치(土壤分析値)(표본(標本)크기 : 2,808개(個))를 이용(利用)하여 다음과 같은 결과(結果)를 얻었다. 1. 토성별(土性別) 평균(平均) 유효수분함량(有效水分含量)은 사토(砂土) 4.7, 양질토양(壤質砂土) 7.7, 사양토(砂壤土) 13.2, 토양(壤土) 17.7, 미사질양토(徵砂質壤土) 19.2, 식양토(埴壤土) 15.9, 사질식양토(砂質埴壤土) 14.5, 미세질양토(徵砂質埴壤土) 18.7, 미사질식토(微砂質埴土) 17.3 그리고 식토(埴土) 14.9%이었다. 2. 포장용수량(圃場容水量)일때의 수분함량(水分含量)과 유효수분함량(有效水分含量)은 대체로 조립질토양(組粒質土壤)에서 모래 그리고 세립질토양(細粒質土壤)에서는 유기물(有機物), 영구위조점(永久萎凋點)에서의 수분함량(水分含量)은 점토(粘土)가 가장 크게 관계(關係)가 있었다. 3. 유효수분함량(有效水分含量)은 포장용수량(圃場容水量)일때의 수분함량(水分含量)과 고도(高度)의 유의성(有意性)을 나타내었고 영구위조점(永久萎凋點)일때의 수분함량(水分含量)과는 상관(相關)이 없었다. 유효수분함량(有效水分含量)은 미사(微砂)의 영향(影響)을 크게 받았다. 4. 토성별(土性別) 포장용수량(圃場容水量)일때의 수분함량(水分含量) 및 유효수분함량(有效水分含量)의 추정식(推定式)이 얻어졌다.

• 대한치과보철학회지
• /
• 제41권2호
• /
• pp.182-206
• /
• 2003

Statement of problem : Successful osseointegration of endosseous threaded implants is dependent on many factors. These may include the surface characteristics and gross geometry of implants, the quality and quantity of bone where implants are placed, and the magnitude and direction of stress in functional occlusion. Therefore clinical quantitative measurement of primary stability at placement and functional state of implant may play a role in prediction of possible clinical symptoms and the renovation of implant geometry, types and surface characteristic according to each patients conditions. Ultimately, it may increase success rate of implants. Purpose : Many available non-invasive techniques used for the clinical measurement of implant stability and osseointegration include percussion, radiography, the $Periotest^{(R)}$, Dental Fine $Tester^{(R)}$ and so on. There is, however, relatively little research undertaken to standardize quantitative measurement of stability of implant and osseointegration due to the various clinical applications performed by each individual operator. Therefore, in order to develop non-invasive experimental method to measure stability of implant quantitatively, the resonance frequency analyzer to measure the natural frequency of specific substance was developed in the procedure of this study. Material & method : To test the stability of the resonance frequency analyzer developed in this study, following methods and materials were used : 1) In-vitro study: the implant was placed in both epoxy resin of which physical properties are similar to the bone stiffness of human and fresh cow rib bone specimen. Then the resonance frequency values of them were measured and analyzed. In an attempt to test the reliability of the data gathered with the resonance frequency analyzer, comparative analysis with the data from the Periotest was conducted. 2) In-vivo study: the implants were inserted into the tibiae of 10 New Zealand rabbits and the resonance frequency value of them with connected abutments at healing time are measured immediately after insertion and gauged every 4 weeks for 16 weeks. Results : Results from these studies were such as follows : The same length implants placed in Hot Melt showed the repetitive resonance frequency values. As the length of abutment increased, the resonance frequency value changed significantly (p<0.01). As the thickness of transducer increased in order of 0.5, 1.0 and 2.0 mm, the resonance frequency value significantly increased (p<0.05). The implants placed in PL-2 and epoxy resin with different exposure degree resulted in the increase of resonance frequency value as the exposure degree of implants and the length of abutment decreased. In comparative experiment based on physical properties, as the thickness of transducer increased, the resonance frequency value increased significantly(p<0.01). As the stiffness of substances where implants were placed increased, and the effective length of implants decreased, the resonance frequencies value increased significantly (p<0.05). In the experiment with cow rib bone specimen, the increase of the length of abutment resulted in significant difference between the results from resonance frequency analyzer and the $Periotest^{(R)}$. There was no difference with significant meaning in the comparison based on the direction of measurement between the resonance frequency value and the $Periotest^{(R)}$ value (p<0.05). In-vivo experiment resulted in repetitive patternes of resonance frequency. As the time elapsed, the resonance frequency value increased significantly with the exception of 4th and 8th week (p<0.05). Conclusion : The development of resonance frequency analyzer is an attempt to standardize the quantitative measurement of stability of implant and osseointegration and compensate for the reliability of data from other non-invasive measuring devices It is considered that further research is needed to improve the efficiency of clinical application of resonance frequency analyzer. In addition, further investigation is warranted on the standardized quantitative analysis of the stability of implant.

• 지능정보연구
• /
• 제26권1호
• /
• pp.135-149
• /
• 2020

인공지능을 기반으로 한 다양한 연구들이 현대사회에 많은 변화를 불러일으키고 있다. 금융시장 역시 예외는 아니다. 로보어드바이저 개발이 활발하게 진행되고 있으며 전통적 방식의 단점을 보완하고 사람이 분석하기 어려운 부분을 대체하고 있다. 로보어드바이저는 인공지능 알고리즘으로 자동화된 투자 결정을 내려 다양한 자산배분 모형과 함께 활용되고 있다. 자산배분 모형 중 리스크패리티는 대표적인 위험 기반 자산배분 모형의 하나로 큰 자산을 운용하는 데 있어 안정성을 나타내고 현업에서 역시 널리 쓰이고 있다. 그리고 XGBoost 모형은 병렬화된 트리 부스팅 기법으로 제한된 메모리 환경에서도 수십억 가지의 예제로 확장이 가능할 뿐만 아니라 기존의 부스팅에 비해 학습속도가 매우 빨라 많은 분야에서 널리 활용되고 있다. 이에 본 연구에서 리스크패리티와 XGBoost를 장점을 결합한 모형을 제안하고자 한다. 기존에 널리 사용되는 최적화 자산배분 모형은 과거 데이터를 기반으로 투자 비중을 추정하기 때문에 과거와 실투자 기간 사이의 추정 오차가 발생하게 된다. 최적화 자산배분 모형은 추정 오차로 인해 포트폴리오 성과에서 악영향을 받게 된다. 본 연구는 XGBoost를 통해 실투자 기간의 변동성을 예측하여 최적화 자산배분 모형의 추정 오차를 줄여 모형의 안정성과 포트폴리오 성과를 개선하고자 한다. 본 연구에서 제시한 모형의 실증 검증을 위해 한국 주식시장의 10개 업종 지수 데이터를 활용하여 2003년부터 2019년까지 총 17년간 주가 자료를 활용하였으며 in-sample 1,000개, out-of-sample 20개씩 Moving-window 방식으로 예측 결과값을 누적하여 총 154회의 리밸런싱이 이루어진 백테스팅 결과를 도출하였다. 본 연구에서 제안한 자산배분 모형은 기계학습을 사용하지 않은 기존의 리스크패리티와 비교하였을 때 누적수익률 및 추정 오차에서 모두 개선된 성과를 보여주었다. 총 누적수익률은 45.748%로 리스크패리티 대비 약 5% 높은 결과를 보였고 추정오차 역시 10개 업종 중 9개에서 감소한 결과를 보였다. 실험 결과를 통해 최적화 자산배분 모형의 추정 오차를 감소시킴으로써 포트폴리오 성과를 개선하였다. 포트폴리오의 추정 오차를 줄이기 위해 모수 추정 방법에 관한 다양한 연구 사례들이 존재한다. 본 연구는 추정 오차를 줄이기 위한 새로운 추정방법으로 기계학습을 제시하여 최근 빠른 속도로 발전하는 금융시장에 맞는 진보된 인공지능형 자산배분 모형을 제시한 점에서 의의가 있다.