• Survey Research
• /
• v.7 no.2
• /
• pp.53-69
• /
• 2006

Weights can be made and imposed in both sample design stage and analysis stage in a sample survey. While in design stage weights are related with sample data acquisition quantities such as sample selection probability and response rate, in analysis stage weights are connected with external quantities, for instance population quantities and some auxiliary information. The final weight is the product of all weights in both stage. In the present paper, we focus on the weight in analysis stage and investigate the effect of such weights imposed on the weighted mean when estimating the population mean. We consider a finite population with a pair of fixed survey value and weight in each unit, and suppose equal selection probability designs. Under the condition we derive the formulas of the bias as well as mean square error of the weighted mean and show that the weighted mean is biased and the direction and amount of the bias can be explained by the correlation between survey variate and weight: if the correlation coefficient is positive, then the weighted mein over-estimates the population mean, on the other hand, if negative, then under-estimates. Also the magnitude of bias is getting larger when the correlation coefficient is getting greater. In addition to theoretical derivation about the weighted mean, we conduct a simulation study to show quantities of the bias and mean square errors numerically. In the simulation, nine weights having correlation coefficient with survey variate from -0.2 to 0.6 are generated and four sample sizes from 100 to 400 are considered and then biases and mean square errors are calculated in each case. As a result, in the case or 400 sample size and 0.55 correlation coefficient, the amount or squared bias of the weighted mean occupies up to 82% among mean square error, which says the weighted mean might be biased very seriously in some cases.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2015.05a
• /
• pp.39-39
• /
• 2015

극치 수문(Hydrologic extremes)분야에서는 수문자료의 분포에 따라 Gumbel, GEV, 그리고 GLO 분포와 같은 다양한 확률통계 분포형이 존재한다. GEV와 GLO 분포형의 경우 Gumbel 분포형과 달리 형상매개변수가 포함된 3변수 분포형으로써 이상 기후 현상으로 인한 잦은 극치 수문사상을 표현하는데 좀 더 유연한 것으로 알려져 있다. 특히 GLO 분포형의 경우 영국에서 홍수빈도해석 시 적정분포형으로 선정된바 있다(Institute of Hydrology, 1999). 다양한 분포형 중에서 표본 자료를 대표할 수 있는 분포형을 선정하는 통계적 기법이 적합도 검정이다. 적합도 검정에는 $x^2$-검정, Cramer von-Mises 검정, Kolmogorov-Smirnov 검정, PPCC(probability plot correlation coefficient, 확률도시 상관계수)검정 등이 있으며 그 중 PPCC 검정은 이용방법이 간편하면서도 뛰어난 기각능력을 보이는 것으로 알려져 있다. 본 연구에서는 극치 수문분야에서 널리 이용되고 있는 GLO 분포형을 대상으로 자료의 왜곡도 영향을 고려할 수 있는 확률도시 상관계수 검정의 검정통계량을 추정하여 보았다.

• Proceedings of the Korean Association for Survey Research Conference
• /
• 2003.06a
• /
• pp.34-41
• /
• 2003

In many cases, the measurement error variances may be functions of the unknown true values or related covariates. This paper develops estimators of the parameters of a linear measurement error variance function based on wi thin-unit sample variaoces. This paper devotes to: (1) define measurement error scale factor $\delta$: (2) develop estimators of the parameters of the 1inear measurement error variance function under stratified multistage sampling design and small error conditions; (3) use propensity methods to adjust survey weights to account for possible selection effects at the replicate level. The proposed methods are applied to medical examination data from the U S Third National Health and Nutrition Examination Survey(NHANES III)

• The Korean Journal of Applied Statistics
• /
• v.28 no.3
• /
• pp.485-494
• /
• 2015

Canonical correspondence analysis is an ordination method used to visualize the relationships among sites, species and environmental variables. However, projection results are fluctuations if the samples slightly change and consistent interpretation on ecological similarity among species tends to be difficult. We use the bootstrap methods for canonical correspondence analysis to solve this problem. The bootstrap method results show that the variations of coordinate points are inversely proportional to the number of observations and coverage rates with bootstrap confidence interval approximates to nominal probabilities.

• Journal of the Korean Data and Information Science Society
• /
• v.26 no.3
• /
• pp.593-610
• /
• 2015

Adaptive Rejection Sampling (ARS) method is a well-known random number generator to acquire a random sample from a probability distribution, and has the advantage of improving the proposal distribution during the sampling procedures, which update it closer to the target distribution. However, the use of ARS is limited since it can be used only for the target distribution in the form of the log-concave function, and thus various methods have been proposed to overcome such a limitation of ARS. In this paper, we attempt to compare five random number generators based on ARS in terms of adequacy and efficiency. Based on empirical analysis using simulations, we discuss their results and make a comparison of five ARS-based methods.

• Journal of the Korean Data Analysis Society
• /
• v.20 no.6
• /
• pp.2883-2893
• /
• 2018

We suggest a stratified conditional unrelated question randomized response model to more efficiently estimate a sensitive character A when the population is composed of several strata. In that model, only the respondents who answered "yes" through randomization device which was consisted of a less sensitive character B and a question forcing to answer "yes" respond to our suggested model and we deal with two allocation problems of proportional allocation and optimal one. We expand the suggested model into two sample stratified conditional unrelated question model to cover the case of unknowing unrelated character and deduce minimal variance through optimal sample size of stratum h. Finally, we show that the suggested model is more efficiency than stratified unrelated models and the stratified Carr et al.'s model (1982) under some given conditions, and show numerically that the smaller the values ${\pi}_{h2}$ and ${\pi}_{hy}$, the more efficiency the fit of the model.

• Journal of Korea Society of Industrial Information Systems
• /
• v.4 no.4
• /
• pp.47-52
• /
• 1999

In this paper, we develop noninformative priors that are used for estimating the reliability of stress-strength system under the Burr-type X distribution. A class of priors is found by matching the coverage probabilities of one-sided Bayesian credible interval with the corresponding frequentist coverage probabilities. It turns out that the reference prior is a first order matching prior. The propriety of posterior under matching prior is provided. The frequentist coverage probabilities are given for small samples.

• KDI Journal of Economic Policy
• /
• v.36 no.3
• /
• pp.1-23
• /
• 2014

This paper has studied the monetary policy in Korea with a time varying VAR model using four key macroeconomic variables. First, inclusion of the exchange rate was a crucial factor in evaluating Korean monetary policy since the monetary policy demonstrated sensitivity to exchange rate movements during the crisis periods of both the Asian financial crisis of 1997 and the global financial crisis of 2008. Second, a specification of the stochastic volatilities in TVP-VAR model is important in explaining excessive movements of all variables in the sample. The overall moderation of variables in 2000s was more or less due to a reduction of the stochastic volatilities but also somewhat due to the macroeconomic fundamental structures captured by impulse response functons. Third, the degree of the monetary policy effectiveness of inflation was mitigated in recent periods but with increased persistence. Lastly, the monetary policy stance towards inflation stabilization has advanced ever since the inflation targeting scheme was adopted. However, there still seems to be a room for improvement in this aspect since the degree of the monetary policy stance towards inflation stabilization was relatively weaker than to output stabilization.

• Proceedings of the Technology Innovation Conference
• /
• 2009.02a
• /
• pp.313-332
• /
• 2009

본 연구에서는 국가의 연구개발활동조사에서 기업연구개발활동 통계에 대한 효과적인 산출방법을 제시하고자 하였다. 이를 위하여 국내 외 연구개발 통계방법을 조사한 후 이를 토대로 우리나라에서 기업연구개발활동에 대한 자료의 수집 및 분석에 대한 개선방안을 제시하였다. 대부분의 국가에서는 대기업은 전수조사, 소규모 기업은 표본조사를 수행하고 있으나, 우리나라에서는 연구소 등록법인에 대하여 전수조사를 행하고 있다. 전수조사는 비용이 많이 들고 비 표본오차로 인하여 모집단에 대한 체계적인 추정이 불가능하다는 문제점이 있다. 현재 산업기술진흥협회에 등록된 연구기관의 수가 20,000개를 넘어서고 있어 전수조사는 한계에 다다른 것으로 생각되어 표본조사 도입에 대한 타당성과 방법론을 중점적으로 검토하였다. 먼저, 표본조사의 타당성을 평가하기 위하여 현재 전수조사를 통해 수집된 자료를 이용하여 표본조사를 수행한 결과를 비교 분석하였다. 산업별(24개), 그룹별(8개)로 구분하여 216개 셀별로 모집단수/표본수를 곱하여 산정 (셀별추정법)한 결과, 전수 통계치와 거의 동일하게 나타났다. 따라서, 산업별, 그룹별로 세분하여 모집단수/표본수를 곱하여 추정하는 셀별추정법이 타당한 것으로 평가할 수 있다. 이상의 분석결과를 토대로 새로운 조사설계방안을 제시하면 다음과 같다. 직전연도 조사기업은 직전연도 연구개발비 수준과 기업종류(대기업, 벤처기업, 중소기업), 그리고 산업에 따라 셀을 분할한다. 대기업, 연구개발비 수준이 높은 기업 등 주요한 셀에 대하여는 전수조사를 실시한다. 나머지 셀에 대하여는 각 셀별 연구개발지출의 분포가 동질적이기 때문에 표본 추출방법은 단순임의추출법(SRS)을 사용한다. 다만 전년도 미계상된(또는 미포함된) 기업에 대하여는 신규 대형 연구소 진입 등을 고려하여 규모비례확률추출법(PPS)을 고려하는 것이 바람직할 것으로 판단된다. 일부 기업들이 특정 항목에 대한 자료를 제공하지 않는 항목무응답의 경우, 누락된 자료에 대하여는 대체기법(Imputation Algorithm)에 따라 이를 추정한다. 이러한 표본조사방법은 전수조사에서 발생하는 비 표본오차를 해소하고, 자료수집비용 및 소규모기업의 행정적 부담을 경감할 수 있다는 장점이 있다. 향후 연구에서는 좀 더 구체적인 조사방법론을 강구할 필요가 있으며, 이와 함께, 연구개발에 대한 다양한 측면의 정보를 수집하기 위해 새로운 설문지를 개발할 필요성이 있다.

• Journal of Korean Society of Forest Science
• /
• v.107 no.4
• /
• pp.402-411
• /
• 2018

This study was conducted to develop ingrowth estimation equations on Pinus densiflora found in Gangwon Province and in the center of Korean Peninsula, based on the National Forest Inventory (NFI)'s permanent sampling plot data. For this study, identical sampling plots in $5^{th}$ and $6^{th}$ NFI data were collected in order to identify ingrowth amounts for the last 5 years. Following two-stage approaches in developing the ingrowth estimation equations, the logistic regression model was used in the first stage to estimate the ingrowth probability. In the second stage, regression analysis on sampling plots with ingrowth occurrence was used to estimate the ingrowth amount. A candidate model was finally selected as an optimal model after a verification based on three evaluation statistics which include mean difference (MD), standard deviation of difference (SDD) and standard error of difference (SED). In results, a logistic regression model based on the number of sampling plot which did not result in ingrowth (model VI), was selected for an ingrowth probability estimation equation and exponential function including the species composition (SC) variable was optimal for an ingrowth estimation equation (model VII). The ingrowth estimation equations developed in this study also evaluated the estimation ability in various forest stand conditions, and no particular issue in fitness or applicability was observed.