• Journal of the Korean Statistical Society
• /
• 제17권1호
• /
• pp.9-23
• /
• 1988

The classicla theory of statistical calibration assumes that the standard measurement is exact. From a realistic point of view, however, this assumption needs to be relaxed so that more meaningful calibration procedures may be developed. This paper presents a model which explicitly considers errors in both standard and nonstandard measurements. Under the assumption that replicated observations are available in the calibration experiment, three estimation techniques (ordinary least squares, grouping least squares, and maximum likelihood estimation) combined with two prediction methods (direct and inverse prediction) are compared in terms of the asymptotic mean square error of prediction.

• 응용통계연구
• /
• 제21권1호
• /
• pp.1-17
• /
• 2008

형제 쌍(sibpair)의 연속형 형질(continuous traits) 자료를 이용한 유전연관성 검정 법(linkage test)으로서 Haseman과 Elston (1972)의 최소제곱(ordinary least square, OLS) 회귀분석법이 주로 사용된다. 비모수적 방법으로서 제시된 Kruglyak과 Lander (1995)의 검정통계량은 Haseman과 Elston (1972)의 방법에 대응되는 방법처럼 보이지만 실제로는 매우 다르다. 본 논문에서는 Kruglyak와 Lander (1995)의 검정통계량과 Haseman과 Elston (1972)의 검정통계량의 관계를 설명하고 모의실험으로 두 검정통계량의 검정력을 비교한다. 유전연관성에 사용되는 형제 자료의 특징은 한정된 설명변수의 값에 매우 많은 자료가 반복(replicated)되었다는 점이며, 이러한 반복 자료에 더욱 적절한 가중 회귀분석법을 제안한다. 가중 회귀분석법의 효율성을 정규분포 또는 정규분포가 아닌 연속형 형질 모의실험 자료로 알아본 결과 형제 쌍 자료의 유전연관성 검정에서 가중 회귀분석법이 다른 검정법들보다도 검정력이 높음을 확인하였다.

• 한국수자원학회:학술대회논문집
• /
• 한국수자원학회 2007년도 학술발표회 논문집
• /
• pp.241-245
• /
• 2007

국내 홍수빈도 분포의 매개변수 추정에서 지점추정(at-site estimate) 방법은 유량 자료의 부족으로 발생하는 표본오차(sampling error)가 크기 때문에 충분한 유량 자료를 보유한 지점에 한하여 제한적으로 사용되고 있다. 대안으로 동질성을 가진 유역의 유량 자료를 모아 지역 매개변수를 추정하는 지수홍수법(Index Flood Method)이 제안되기도 하였으나, 이질성이 큰 우리나라의 유역특성 때문에 적용이 쉽지 않다. Stedinger와 Tasker가 1986년 제안한 GLS(Generalized Least Square) 기법은 유역을 동질지역으로 구분할 필요가 없으며 지점들간의 상관관계와 이분산성을 고려할 수 있어, 국내 홍수빈도 해석을 위해서 꼭 도입해야할 기법으로 생각된다. 본 연구에서는 기존의 GLS 기법의 단점을 보완한 Bayesian-GLS 기법을 이용하여, 국내 대유역에 골고루 위치하며 댐의 영향을 받지 않는 31개 지점의 연최대 일유량 시계열의 L-변동계수(L-moment coefficient variation)와 L-왜도계수(L-moment coefficient skewness)를 추정할 수 있는 회귀모형을 제안하였다. 위 회귀모형을 구성하기 위한 유역특성으로는 유역면적, 유역경사, 유역평균강우 등을 사용하였다. Bayesian-GLS (B-GLS) 적용 결과를 OLS(Ordinary Least Square) 및 Bayesian-GLS 기법에서 지점간의 상관관계를 고려하지 않는 Bayesian-WLS(Weighted Least Square)와 비교 평가하여 그 우수성을 입증하였다. 따라서 본 연구에서 제안된 B-GLS에 의한 지역회귀모형은 국내의 미계측유역이나 또는 관측 길이가 짧은 계측유역의 홍수빈도분석을 위해 매우 유용할 것으로 기대된다.년 홍수 피해가 발생하고 있지만, 다른 한편 인구밀도가 높고 1인당 가용 수자원이 상대적으로 적기 때문에 국지적 물 부족 문제를 경험하고 있다. 최근 국제적으로도 농업용수의 물 낭비 최소화와 절약 노력 및 타 분야 물 수요 증대에 대한 대응 능력 제고가 매우 중요한 과제로 부각되고 있다. 2006년 3월 멕시코에서 개최된 제4차 세계 물 포럼에서 국제 강 네트워크는 "세계 물 위기의 주범은 농경지", "농민들은 모든 물 위기 논의에서 핵심"이라고 주장하고, 전 프랑스 총리 미셀 로카르는 "...관개시설에 큰 문제점이 있고 덜 조방적 농업을 하도록 농민들을 설득해야 한다. 이는 전체 농경법을 바꾸는 문제..."(segye.com, 2006. 3. 19)라고 주장하는 등 세계 물 문제 해결을 위해서는 농업용수의 효율적 이용 관리가 중요함을 강조하였다. 본 연구는 이러한 국내외 여건 및 정책 환경 변화에 적극적으로 대처하고 물 분쟁에 따른 갈등해소 전략 수립과 효율적인 물 배분 및 이용을 위한 기초연구로서 농업용수 수리권과 관련된 법 및 제도를 분석하였다.. 삼요소의 시용 시험결과 그 적량은 10a당 질소 10kg, 인산 5kg, 및 가리 6kg 정도였으며 질소는 8kg 이상의 경우에는 분시할수록 비효가 높았으며 특히 벼의 후기 중점시비에 의하여 1수영화수와 결실율의 증대가 크게 이루어졌다. 3. 파종기와 파종량에 관한 시험결과는 공시품종선단의 파종적기는 4월 25일부터 5월 10일경까지 인데 이 기간중 일찍 파종하는 경우에 파종적량은 10a당 약 8${\ell}$이고 늦은 경우에는 12${\ell}$ 정도였다. 여기서 늦게 파종한 경우 감수의 가장 큰 원인은 1수영화수가 적어지기 때문이었다. 4. 건답직파에 대한 담수상태로 관수를 시작하는 적기는 파종후

• Journal of Electrical Engineering and Technology
• /
• 제7권4호
• /
• pp.636-645
• /
• 2012

This study introduces an information granular-based fuzzy radial basis function neural networks (FRBFNN) based on multiobjective optimization and weighted least square (WLS). An improved multiobjective space search algorithm (IMSSA) is proposed to optimize the FRBFNN. In the design of FRBFNN, the premise part of the rules is constructed with the aid of Fuzzy C-Means (FCM) clustering while the consequent part of the fuzzy rules is developed by using four types of polynomials, namely constant, linear, quadratic, and modified quadratic. Information granulation realized with C-Means clustering helps determine the initial values of the apex parameters of the membership function of the fuzzy neural network. To enhance the flexibility of neural network, we use the WLS learning to estimate the coefficients of the polynomials. In comparison with ordinary least square commonly used in the design of fuzzy radial basis function neural networks, WLS could come with a different type of the local model in each rule when dealing with the FRBFNN. Since the performance of the FRBFNN model is directly affected by some parameters such as e.g., the fuzzification coefficient used in the FCM, the number of rules and the orders of the polynomials present in the consequent parts of the rules, we carry out both structural as well as parametric optimization of the network. The proposed IMSSA that aims at the simultaneous minimization of complexity and the maximization of accuracy is exploited here to optimize the parameters of the model. Experimental results illustrate that the proposed neural network leads to better performance in comparison with some existing neurofuzzy models encountered in the literature.

• 한국측량학회지
• /
• 제35권1호
• /
• pp.31-39
• /
• 2017

The goal of this study was to efficiently analyze the relationships of the number of thefts with related factors, considering the spatial patterns of theft crimes. Theft crime data for a 5-year period (2009-2013) were collected from Haeundae Police Station. A logarithmic transformation was performed to ensure an effective statistical analysis and the number of theft crimes was used as the dependent variable. Related factors were selected through a literature review and divided into social, environmental, and defensive factors. Seven factors, were selected as independent variables: the numbers of foreigners, aged persons, single households, companies, entertainment venues, community security centers, and CCTV (Closed-Circuit Television) systems. OLS (Ordinary Least Squares) and GWR (Geographically Weighted Regression) were used to analyze the relationship between the dependent variable and independent variables. In the GWR results, each independent variable had regression coefficients that differed by location over the study area. The GWR model calculated local values for, and could explain the relationships between, variables more efficiently than the OLS model. Additionally, the adjusted R square value of the GWR model was 10% higher than that of the OLS model, and the GWR model produced a AICc (Corrected Akaike Information Criterion) value that was lower by 230, as well as lower Moran's I values. From these results, it was concluded that the GWR model was more robust in explaining the relationship between the number of thefts and the factors related to theft crime.

• The Journal of Asian Finance, Economics and Business
• /
• 제7권10호
• /
• pp.675-685
• /
• 2020

The main purpose of this study is to investigate the potential effect of ownership concentration on the relationship between board composition and bank performance. The study employs a sample of Saudi banks listed on Saudi stock exchange (TADAUWL) over the period from 2011 to 2018. To test the study hypotheses and control for endogeneity issues, the Ordinary Least Square (OLS) and the Two-Stage Least Squares (2SLS) techniques are used. The empirical results reveal a significant negative moderating effect of ownership concentration on the association between board composition and bank performance, which confirms the study argument and supports hypotheses. The results indicate that board composition in terms of independent board members, executive board members, and non-executive board members in banks with higher ownership concentration have a weaker positive influence on bank performance. For control variables, the results are almost consistent with theoretical perspectives and previous empirical evidence. The results of this study have important implications for regulatory authorities, companies, and market participants in Saudi Arabia and countries with high concentrated ownership to understand how ownership concentration could affect corporate governance and firm performance and to identify appropriate actions to protect board composition from the influence of ownership concentration.

• Communications for Statistical Applications and Methods
• /
• 제24권3호
• /
• pp.303-315
• /
• 2017

Yoo (2015, Statistics and Probability Letters, 99, 109-113) derives theoretical results in an optimal sufficient dimension reduction with singular inner-product matrix. The results are promising, but Yoo (2015) only presents one simulation study. So, an evaluation of its practical usefulness is necessary based on numerical studies. This paper studies the asymptotic behaviors of Yoo (2015) through various simulation models and presents a real data example that focuses on ordinary least squares. Intensive numerical studies show that the $x^2$ test by Yoo (2015) outperforms the existing optimal sufficient dimension reduction method. The basis estimation by the former can be theoretically sub-optimal; however, there are no notable differences from that by the latter. This investigation confirms the practical usefulness of Yoo (2015).

• 보건행정학회지
• /
• 제5권2호
• /
• pp.18-34
• /
• 1995

To clarify the relationship between the medical supply(medical persons and goods) and the use of bed, the author has made comparison among OECD 24 countries. Per Capita Bed-days can be divided into Average Length of Stay and Admission Rate, and these three variables were regressed upon both In-patient Care Beds of all medical institutions including acute somatic, psychiatric, special, nursing homes and other long-term care and Share of Total Health Employment in Total Employment. The result of regression analysis shows a statistically significant positive relationship between In-patient Care Beds and Average Length of Stay, and negative relationship between Share of Total Health Employment and Admission Rate. In addition to Ordinary Least Square(OLS) estimation, amended Bounded Influence Estimation(BIE) was also made to adjust the influence of outliers. Japan shows a very large number of In-patient Care Beds and a very low Share of Total Health Employment, and this medical situation is judged to have close relation to her long Average Length of Stay and low Admission Rate.

• 한국소음진동공학회논문집
• /
• 제16권7호
• /
• pp.769-776
• /
• 2006

A measured frequency response function can be represented as a ratio of two polynomials. A curve-fitting of frequency responses with Legendre polynomialis suggested in the paper. And the suggested curve-fitting algorithm is based on the least-square error method. Since the Legendre polynomials satisfy the orthogonality condition, the curve-fitting with the polynomials results to more reliable curve-fitting than ordinary polynomial method. Though the proposed curve-fitting with Legendre polynomials cannot cover all frequency range of interest, example shows that the suggested method is quite applicable in a limited frequency band.

• Communications for Statistical Applications and Methods
• /
• 제1권1호
• /
• pp.102-111
• /
• 1994

We Propose a method of sensitivity analysis in latent root regression analysis (LRRA). For this purpose we derive the quantities ${\beta\limits^\wedge \;_{LRR}}^{(1)}$, which correspond to the theoretical influence function $I(x, y \;;\;\beta\limits^\wedge \;_{LRR})$ for the regression coefficient ${\beta\limits^\wedge}_{LRR}$ based on LRRA. We give a numerical example for illustration and also investigate numerically the relationship between the estimated values of ${\beta\limits^\wedge \;_{LRR}}^{(1)}$ with the values of the other measures called sample influence curve(SIC) based on the recomputation for the data with a single observation deleted. We also discuss the comparision among the results of LRRA, ordinary least square regression analysis (OLSRA) and ridge regression analysis(RRA).