• 응용통계연구
• /
• 제14권1호
• /
• pp.91-101
• /
• 2001

다중나무예측치들이 한 개의 나무 예측치 보다 검증용 자료 오분류률을 줄이는데 있어서 더 정확하다 라는 것은 잘 알려져 있는 사실이다. 다중나무를 생성하는 두 가지 방법이 있다. 하나는 원래의 훈련용 자료를 재 추출하여 수정된 훈련용자료들을 만든 다음에 각각의 수정된 훈련용 자료에 근거하여 나무를 만드는 것이다. arcing 알고리즘이 효율적이라고 알려져있다. 다른 방법은 각각의 마디에서 최적 분리의 후보들 중에서 랜덤하게 하나를 선택하여 나무를 생성하는데에, 이 과정을 반복하면 원래의 훈련용 자료에 대해서 비교적 좋은 나무들을 생성하리라 기대되다. 우리는 arcing의 각 단계에서 후자의 다중회귀나무예측치들을 사용하는 결합다중회귀나무예측치를 제안하고, 효능 있는 화합물들을 찾기 위한 고속의 대량 선별 자료 분석의 예를 통해서 예측방법들의 효율성을 비교한다.

• 한국대기환경학회지
• /
• 제13권1호
• /
• pp.79-89
• /
• 1997

In order to reduce the outbreaks of short-term high concentrations and its impacts, we developed the models which predicted tomorrow's maximum hourly concentrations of $O_3$, TSP, SO$_2$, NO$_2$ and CO. Statistical methods like multi regressions were used because it must be operated easily under the present conditions. 47 independent variables were used, which included observed concentrations of air pollutants, observed and forcasted meteorological data in 1994 at Seoul and its surrounding areas. We subdivided Seoul into 4 areas coinciding with the present ozone warning areas. 4 kinds of seasonal models were developed due to the seasonal variations of observed concentrations, and 2 kinds of data models for the unavailable case of forecasted meteorological data. By comparing the $R^2$and root mean square error(hearafter 'RMSE') of each model, we confirmed that the models including forecasted data showed higher accuracy than ones using observed only. It was also shown that the higher the seasonal mean concentrations, the larger the RMSE. There was no distinct difference between the results of 4 areal models. In case of test run using 1995's data, the models predicted well the trends of daily variation of concentrations and the days when the possibility of outbreak of high concentarion was high. This study showed that it was reasonable to use those models as operational ones, because the $R^2$ and RMSE of models were smaller than those of operational/research models such as in South Coast Air Basin, CA, USA.

• Structural Engineering and Mechanics
• /
• 제74권6호
• /
• pp.821-832
• /
• 2020

The fundamental period is an important parameter for seismic design and seismic risk assessment of building structures. In this paper, a simplified theoretical method to predict the fundamental period of masonry infilled reinforced concrete (RC) frame is developed based on the basic theory of engineering mechanics. The different configurations of the RC frame as well as masonry walls were taken into account in the developed method. The fundamental period of the infilled structure is calculated according to the integration of the lateral stiffness of the RC frame and masonry walls along the height. A correction coefficient is considered to control the error for the period estimation, and it is determined according to the multiple linear regression analysis. The corrected formula is verified by shaking table tests on two masonry infilled RC frame models, and the errors between the estimated and test period are 2.3% and 23.2%. Finally, a probability-based method is proposed for the corrected formula, and it allows the structural engineers to select an appropriate fundamental period with a certain safety redundancy. The proposed method can be quickly and flexibly used for prediction, and it can be hand-calculated and easily understood. Thus it would be a good choice in determining the fundamental period of RC frames infilled with masonry wall structures in engineering practice instead of the existing methods.

• 한국농공학회지
• /
• 제44권5호
• /
• pp.41-53
• /
• 2002

This study was conducted to derive the regional design rainfall by the regional frequency analysis based on the regionalization of the precipitation suggested by the first report of this project. According to the regions and consecutive durations, optimal design rainfalls were derived by the regional frequency analysis for L-moment in the second report of this project. Using the LH-moment ratios and Kolmogorov-Smirnov test, the optimal regional probability distribution was identified to be the Generalized extreme value (GEV) distribution among applied distributions. regional and at-site parameters of the GEV distribution were estimated by the linear combination of the higher probability weighted moments, LH-moment. Design rainfall using LH-moments following the consecutive duration were derived by the regional and at-site analysis using the observed and simulated data resulted from Monte Carlo techniques. Relative root-mean-square error (RRMSE), relative bias (RBIAS) and relative reduction (RR) in RRMSE for the design rainfall were computed and compared in the regional and at-site frequency analysis. Consequently, it was shown that the regional analysis can substantially more reduce the RRMSE, RBIAS and RR in RRMSE than at-site analysis in the prediction of design rainfall. Relative efficiency (RE) for an optimal order of L-moments was also computed by the methods of L, L1, L2, L3 and L4-moments for GEV distribution. It was found that the method of L-moments is more effective than the others for getting optimal design rainfall according to the regions and consecutive durations in the regional frequency analysis. Diagrams for the design rainfall derived by the regional frequency analysis using L-moments were drawn according to the regions and consecutive durations by GIS techniques.

• 한국농공학회지
• /
• 제31권4호
• /
• pp.111-122
• /
• 1989

A multiple box model which is suitable for the prediction of water quality in shallow lakes with active mixing is a water quality model expected to be used widely in estuary reservoir. In this study, a multiple box water quality model for estuary reservoirs (MBQER) was developed arid the applicability of the MBQER was tested by applying data obtained from Asan-estuary reservoir. The results of this study can be summarized as follows. 1. The MBQER, dynamic water quality model, was developed to estimate 10-day water qualities of estuary reservoirs. For the proper analysis and the application of hydraulics needed to build a model, lake hydraulics was simplified by condisering only hydrological inflow and lake mixing currents. The box division in the MBQER is longitudinal one dimension for upper and middle part, and two layers for lower part of the reservoir. 2. The methods of box division for the multiple box model were ekamined and applied to Asan-estuary reservoir. For determining the number of boxes, Pe number and Pk number were used. In case of three boxes, the error by the model simplification would be estimated about 5 % Therefore, in Asan reservoir, the proper number of boxes was three. 3. The MBQER was calibrated and verified using measured data in Asan-estuary reservoir from 1986 to 1988. The Root Mean Squares(RMS) for the differences between measured data and simulated results by the MBQER were 1.10$^{\circ}$C C for water temperature, 75.8mg/1 for salinity, 0.082mg/1 for total-phosphorus showing good estimations. 4. Through the simulation of water temperature and salinity by the MBQER, the exchange flow and the mixing coefficients for the estuary lake were determined. As a result of simulation, the horizontal mixing coefficients in Asan-estuary reservoir were in the range of 1.07X 105 to 1.12X 105 cm$^2$/sec and vertical mixing coefficient was 2.90X 10-1 cm$^2$/sec.

• Nutrition Research and Practice
• /
• 제12권4호
• /
• pp.283-290
• /
• 2018

BACKGROUND/OBJECTIVES: Indirect calorimetry is the gold-standard method for the measurement of resting energy expenditure. However, this method is time consuming, expensive, and requires highly trained personnel. To overcome these limitations, various predictive equations have been developed. The objective of this study was to assess the validity of predictive equations for resting energy expenditure (REE) in Korean non-obese adults. SUBJECTS/METHODS: The present study involved 109 participants (54 men and 55 women) aged between 20 and 64 years. The REE was measured by indirect calorimetry. Nineteen REE equations were evaluated for validity, by comparing predicted and measured REE results. Predictive equation accuracy was assessed by determining percent bias, root mean squared prediction error (RMSE), and percentage of accurate predictions. RESULTS: The measured REE was significantly higher in men than in women (P < 0.001), but the difference was not significant after adjusting for body weight (P > 0.05). The equation developed in this study had an accuracy rate of 71%, a bias of 0%, and an RMSE of 155 kcal/day. Among published equations, the $FAO_{weight}$ equation gave the highest accuracy rate (70%), along with a bias of -4.4% and an RMSE of 184 kcal/day. CONCLUSIONS: The newly developed equation provided the best accuracy in predicting REE for Korean non-obese adults. Among the previously published equations, the $FAO_{weight}$ equation showed the highest overall accuracy. Regardless, at an individual level, the equations could lead to inaccuracies in a considerable number of subjects.

• 한국융합학회논문지
• /
• 제8권11호
• /
• pp.77-83
• /
• 2017

효율적인 예측 분석을 위해서는 문제를 스스로 인식하고 진단하여 시스템이 자율적으로 복구가 가능한 자가 치유 연구가 필요하다. 그러나, 소프트웨어를 개발하는데 있어서 외부상황에 따른 정형화된 컨텍스트 정보 분석 및 적절한 표현 구조를 제시하지 못한다. 본 논문에서는 새로운 목표 시나리오를 기반으로 행위 요소, 데이터, 트랜잭션이 가능한 기능들에 대해 추출 규칙을 적용하여 변경 내용에 따른 예측 분석 방법을 제안한다. 그리고, 요구사항 목표 달성을 위한 성과지표를 통해 예측 분석 내용이 얼마나 부합되었는지 평가하였다. 제안한 방법이 기존 방법들에 비해 성과측정을 통한 부합 결과는 최고 32.8% 높았고, 이에 따른 오차율은 28.9%, 변경 코드는 최고 45.8%가 감소되었다. 이는 목표 시나리오 기반 컨텍스트 규칙을 통해 서비스 가능한 형태로 가공할 수 있음을 보여주며, 문제 발생에 대한 변경 내용을 예측 분석을 통한 성능의 확장이 가능함을 보여준다.

• 방송공학회논문지
• /
• 제20권3호
• /
• pp.449-463
• /
• 2015

동영상에서의 객체 추적은 보안, 색인 및 검색, 감시, 통신, 압축 등 다양한 분야에서 중요하다. 본 논문은 HEVC 비트스트림 상에서의 객체 추적 방법을 제안한다. 복호화를 수행하지 않고, 비트스트림 상에 존재하는 움직임 벡터(MV : Motion Vector)와 부호화 크기 정보를 Spatio-Temporal Markov Random Fields (ST-MRF) 모델에 적용해 객체 움직임의 공간적 및 시간적 특성을 반영한다. 변환계수를 특징점으로 활용하는 객체형태 조정 알고리즘을 적용해 ST-MRF 모델 기반 객체 추적방법에서 나타나는 과분할에 의한 오차전파 문제를 해결한다. 제안하는 방법의 추적성능은 정확도 86.4%, 재현율 79.8%, F-measure 81.1%로 기존방법 대비 평균 F-measure는 약 0.2% 향상하지만 기존방법에서 과분할 및 오차전파가 두드러지는 영상에 대해서는 최대 9% 정도의 성능향상을 보인다. 전체 수행시간은 프레임 당 평균 5.4ms이며 실시간 추적이 가능하다.

• ETRI Journal
• /
• 제36권1호
• /
• pp.134-140
• /
• 2014

Many reports have shown that the access pattern for geospatial tiles follows Zipf's law and that its parameter ${\alpha}$ represents the access characteristics. However, visits to geospatial tiles have temporal and spatial popularities, and the ${\alpha}$-value changes as they change. We construct a mathematical model to simulate the user's access behavior by studying the attributes of frequently visited tile objects to determine parameter estimation algorithms. Because the least squares (LS) method in common use cannot obtain an exact ${\alpha}$-value and does not provide a suitable fit to data for frequently visited tiles, we present a new approach, which uses a moment method of estimation to obtain the value of ${\alpha}$ when ${\alpha}$ is close to 1. When ${\alpha}$ is further away from 1, the method uses the associated cache hit ratio for tile access and uses an LS method based on a critical cache size to estimate the value of ${\alpha}$. The decrease in the estimation error is presented and discussed in the section on experiment results. This new method, which provides a more accurate estimate of ${\alpha}$ than earlier methods, promises more effective prediction of requests for frequently accessed tiles for better caching and load balancing.

• Advances in Computational Design
• /
• 제5권2호
• /
• pp.195-207
• /
• 2020

The present study focuses on the application of artificial neural network (ANN) and Multiple linear Regression (MLR) analysis for developing a model to predict the unconfined compressive strength (UCS) and split tensile strength (STS) of the fiber reinforced clay stabilized with grass ash, fly ash and lime. Unconfined compressive strength and Split tensile strength are the nonlinear functions and becomes difficult for developing a predicting model. Artificial neural networks are the efficient tools for predicting models possessing non linearity and are used in the present study along with regression analysis for predicting both UCS and STS. The data required for the model was obtained by systematic experiments performed on only Kaolin clay, clay mixed with varying percentages of fly ash, grass ash, polypropylene fibers and lime as between 10-20%, 1-4%, 0-1.5% and 0-8% respectively. Further, the optimum values of the various stabilizing materials were determined from the experiments. The effect of stabilization is observed by performing compaction tests, split tensile tests and unconfined compression tests. ANN models are trained using the inputs and targets obtained from the experiments. Performance of ANN and Regression analysis is checked with statistical error of correlation coefficient (R) and both the methods predict the UCS and STS values quite well; but it is observed that ANN can predict both the values of UCS as well as STS simultaneously whereas MLR predicts the values separately. It is also observed that only STS values can be predicted efficiently by MLR.