Earticle

수문기상자료의 빈도해석은 풍수해에 따른 대응 및 시설물의 설계기준에 있어 중요한 요소 중 하나이다. 일반적으로 수문기상자료에 대한 빈도해석의 경우 관측자료는 통계적으로 정상성을 가진다고 가정하고, 확률분포의 매개변수를 고려하는 매개변수적방법을 적용하고 있다. 이러한, 매개변수적 빈도해석을 위해서는 신뢰성 있는 충분한 자료의 수집이 필요하지만, 강수량과다르게 적설량의 경우 계절적 특성과 함께 최근에는 기후변화로 인한 적설량 관측일수 및 평균 최심신적설량이 감소하기때문에 부족한 자료에 대한 문제점을 보완할 필요가 있다. 이에 본 연구에서는 매개변수 빈도해석 방법과 부족한 자료의문제점을 보완할 수 있는 표본 재추출 기법인 Bootstrap방법과 SIR(Sampling Importance Resampling)알고리즘을 적용하여적설량의 빈도해석을 실시하였다. 58개 기상관측소에 대해 재추출된 일 최대 최심신적설량 자료를 이용한 비매개변수적빈도해석을 통해 확률적설량을 산정하고 이를 비교 분석하였다. 빈도별 확률적설량의 증감률을 검토한 결과 매개변수적빈도해석과 비매개변수적 빈도해석에서 증감률을 나타내는 지점들이 대부분 일치하는 것으로 나타났다. 확률적설량은 관측자료와 Bootstrap방법에서 –19.2%∼3.9%, Bootstrap방법과 SIR알고리즘에서 –7.7%∼137.8% 정도의 차이를 보였다. 표본재추출 기법은 관측표본이 적은 적설량의 빈도해석 및 불확실성 범위의 제시가 가능함을 확인할 수 있었고, 이는 여름철태풍과 같이 계절적 특성을 지닌 다른 자연재난의 해석에도 적용될 수 있을 것으로 판단된다.

The frequency analysis of hydrometeorological data is one of the most important factors in response to natural disaster damage, and design standards for a disaster prevention facilities. In case of frequency analysis of hydrometeorological data, it assumes that observation data have statistical stationarity, and a parametric method considering the parameter of probability distribution is applied. For a parametric method, it is necessary to sufficiently collect reliable data; however, snowfall observations are needed to compensate for insufficient data in Korea, because of reducing the number of days for snowfall observations and mean maximum daily snowfall depth due to climate change. In this study, we conducted the frequency analysis for snowfall using the Bootstrap method and SIR algorithm which are the resampling methods that can overcome the problems of insufficient data. For the 58 meteorological stations distributed evenly in Korea, the probability of snowfall depth was estimated by non-parametric frequency analysis using the maximum daily snowfall depth data. The results of frequency based snowfall depth show that most stations representing the rate of change were found to be consistent in both parametric and non-parametric frequency analysis. According to the results, observed data and Bootstrap method showed a difference of –19.2% to 3.9%, and the Bootstrap method and SIR(Sampling Importance Resampling) algorithm showed a difference of –7.7 to 137.8%. This study shows that the resampling methods can do the frequency analysis of the snowfall depth that has insufficient observed samples, which can be applied to interpretation of other natural disasters such as summer typhoons with seasonal characteristics.

P300 숨긴정보검사에서는 관련자극에 대한 P300 진폭이 무관련자극에 대한 P300 진폭보다 더 큰지를 평가한다. 그런데 무관련자극의 시행수가 관련자극의 시행수보다 훨씬 더 크기 때문에 관련자극에 대한 P300 진폭이 과대추정된다는 문제점이 있다. Rosenfeld 등(2008)은 이 문제에 대처하기 위하여 무관련자극의 부트스트랩 표본크기를 관련자극의 표본크기로 축소하여 사용하였다. 일반적으로 부트스트랩 표본크기는 원래의 표본크기와 동일해야만 하며, 부트스트랩 표본크기가 원래의 표본크기보다 작으면 1종 오류율이 유의수준보다 작아지는 문제가 발생한다. 본 연구의 목적은 몬테카를로 연구를 통하여 무관련자극의 부트스트랩 표본크기를 축소하는 수정된 부트스트랩 방법의 1종 오류율을 평가하고, 이러한 오류가 교정될 수 있는지 확인하는 것이다. 실험 1의 결과, 수정된 부트스트랩 방법의 1종 오류율은 약 .073으로 유의수준 .10보다 낮았다. 표준오차를 이용하여 유의수준을 교정한 부트스트랩 방법의 1종 오류율은 약 .140으로 유의수준 .10보다 더 높게 나타나, 수정된 부트스트랩 방법의 오류가 교정되지 않았다. 실험 2에서 수정된 부트스트랩 방법의 오류가 교정되지 않는 이유를 평가하기 위하여 숫자를 이용한 몬테카를로 연구를 수행하였다. 연구결과, 수정된 부트스트랩 방법의 1종 오류율은 약 .054로 유의수준 .10보다 작았으며, 교정된 부트스트랩 방법의 1종 오류율은 약 .10으로 유의수준과 동일하였다. 따라서 수정된 부트스트랩 방법의 오류가 교정되지 않는 이유는 뇌파자료의 특수성 때문인 것으로 나타났다. 이러한 오류를 극복하는 방법에 대해서 논의하였다.

It is evaluated whether the P300 amplitude for the probe is greater than the P300 amplitude for the irrelevant in the P300 concealed information test. However, there is a problem that the P300 amplitude for the probe is overestimated because the number of trials of the irrelevant is much larger than that of the probe. Rosenfeld et al. (2008) attempted to solve this problem by reducing the bootstrap sample size of the irrelevant to the sample size of the probe. In general, the bootstrap sample size must be the same as the original sample size and the type 1 error rate becomes smaller than the significance level if the bootstrap sample size is smaller than the original sample size. The purpose of this study is to evaluate the type 1 error rate of the modified bootstrap method that reduces the bootstrap sample size of irrelevant through Monte Carlo studies and to check whether this error can be corrected. As a result of experiment 1, the type 1 error rate of the modified bootstrap method was about .073, which was lower than the significance level .10. The type 1 error rate of the adjusted bootstrap method with corrected the significance level using the standard error was about .140 which was higher than the significance level .10. Consequently, the error of the modified bootstrap method was not corrected. In order to investigate the reason why the error of the modified bootstrap method was not corrected, a Monte Carlo study using numbers was performed. In the results of experiment 2, the type 1 error rate of the modified bootstrap method was about .054, which was less than the significance level .10, and that of the adjusted bootstrap method was about .10, which was the same as the significance level. It was found that the reason why the error of the modified bootstrap method is not corrected was due to the specificity of the EEG data. The reasons why these errors are not corrected and how to solve these errors were discussed.

본 연구는 90년대 후반 외환위기와 함께 겪은 은행산업의 대규모 구조조정의 효과를 효율성 측면에서 실증분석하여 최근 논의되고 있는 은행의 대형화와 겸업화에 대한 효과를 분석한다. 특히 효율성 요인에 관한 선행연구에서 이용한 자료포락방법(data envelopment analysis)의 계량경제학적 문제점을 해결하기 위하여 Censored Tobit모형과 부트스트랩방법을 이용하여 기존의 연구방법의 추정과 추론상의 문제점을 살펴보았다. 결과적으로 기존 연구방법을 이용할 경우 유의하게 나타난 공적자금비율과 합병여부는 유의하지 않게 나타났으며 이는 기존 연구 결과를 이용한 정책 평가에 주의가 필요함을 보여준다.

This paper investigates the effects of restructuring Korean banking industry following the Asian financial crisis on the economies of scale and scope based on efficiency. The present analysis takes into account the censored dependent variable and the serial correlation that have been ignored in the previous research. The result shows that the proportion of public investment fund and the merger and acquisition, which appeared significant in previous studies, do not have significant effects on the efficiency, but BIS is the only factor that has a significant effect. The study suggests that the policy implications in the literature should be applied with a special care.

본 연구에서는 수문자료와 수문해석과정에 포함되어 있는 불확실성을 고려함으로써 현재 하천 시설물에 적용되고 있는 안전율 혹은 여유고에서 고려하고 있는 불확실성과 비교 검토하였다. 불확실성을 고려하기 위하여 연최대 강우량자료를Bootstrap 방법, SIR 알고리즘을 통해 재추출함으로써 불확실성을 고려한 확률강우량을 산정하였다. 산정된 현재의 확률강우량과 불확실성을 고려한 경우의 확률강우량을 이용하여 HEC-HMS를 통해 홍수량을 산정, HEC-RAS를 통해 지점별 홍수위를 산정하였다. 예상홍수피해액의 산정은 다차원홍수피해산정방법 (MD-FDA)을 이용하였다. 그 결과 SIR 알고리즘을 이용한 경우 예상홍수피해액의 최대값은 현재의 확률강우량을 이용한 경우의 1.22배, Bootstrap 방법에 의한 최대값은 0.92배의 값을 보였다. SIR 알고리즘은 자료의 재추출시 가중치 부여를 위한 우도함수의 영향을 크게 받으며, 이로 인한 불확실성의 구간이 커져 가장 큰 예상홍수 피해액이 도출되었다. 따라서 자료의 적정 위치에 우도를 결정하는 것이 매우 중요하게 작용하는 것을 알 수 있었다.

We estimated the expected flood damage considering uncertainty which is involved in hydrologic processes and data. Actually, this uncertainty represents a freeboard or safety factor in the design of hydraulic structures. The uncertainty was analyzed using Bootstrap method, and SIR algorithm then the frequency based rainfalls were estimated for each method of uncertainty analysis. Also the benefits for each uncertainty analysis were estimated using 'multi-dimensional flood damage analysis(MD-FDA). As a result, the expected flood damage with SIR algorithm was 1.22 times of present status and Boostrap 0.92 times. However when we used SIR algorithm, the likelihood function should be selected with caution for the estimation of the expected flood damage.

P300 숨긴정보검사에서는 조사대상자가 거짓을 말하고 있는지 판단하기 위하여 관련자극의 P300 진폭이 무관련자극의 P300 진폭보다 통계적으로 유의하게 더 큰지를 평가한다. 구체적인 통계적 방법으로 독립표본 t 검증 또는 부트스트랩 방법을 사용할 수 있다. Rosenfeld와 Soskins, Bosh, Ryan(2004)은 “개인 내에서 관련자극과 무관련자극의 P300 평균을 비교하는데 사용하기에는 t 검증이 너무 둔감하다.”면서 부트스트랩 방법을 사용하였다. 본 연구의 목적은 P300 숨긴정보검사에서 t 검증의 검증력이 부트스트랩 방법보다 더 낮은지 평가하는 것이다. 이를 위하여 39명의 실험참가자로부터 측정한 뇌파자료를 이용하여 Monte Carlo 연구를 수행하였다. 연구결과, t 검증과 백분위를 이용한 부트스트랩 방법의 1 종 오류율은 서로 비슷하였으며, 백분위를 이용한 부트스트랩 방법의 검증력이 t 검증의 검증력보다 약간 더 높았다. 두 검증 방법의 1 종 오류율은 모두 유의수준보다 낮은 값을 보였으며, 검증력은 이론적인 t 검증의 검증력보다 약간 낮은 값을 보였다. 반면에 표준오차를 이용한 부트스트랩 방법의 1 종 오류율과 검증력은 이론적인 t 검증의 1 종 오류율 및 검증력과 비슷한 값을 보였으며, t 검증의 검증력보다 실험조건에 따라 .012 ~ .081 더 높았다.

In P300-based concealed information test (P300 CIT), it evaluates whether the P300 amplitude for the probe is significantly greater than that of the irrelevant to determine if the suspect is telling a lie. An independent sample t-test or a bootstrap method can be used as a statistical test to make that decision. Rosenfeld et al. (2004) used the bootstrap method, claiming that “t tests on single sweeps are too insensitive to use to compare mean probe and irrelevant P300s within individuals” and their method has been accepted to date. The purpose of the study is to evaluate whether the power of t-test is lower than that of the bootstrap method in the P300 CIT. The Monte Carlo study was conducted by using EEG collected from 39 participants. The results showed that the type I error rates of the t-test and the percentile bootstrap method were similar and the power of the percentile bootstrap method was slightly higher than that of the t-test. The type I error rates of the t-test and the percentile bootstrap method were slightly lower than the significance level and the powers of the two tests were also slightly lower than that of the theoretical t-test. On the other hand, the type I error rate and power of the standard error Bootstrap method were the same as those of the theoretical t-test and its power was .012 ~ .081 higher than that of t-test depending on experimental conditions.

본 연구는 붓스트랩(bootstrap) 기법을 적용해 축구 스타 휴먼 브랜드 효과를 검증하는데 목적이 있다. 구체적 으로 붓스트랩의 BC법을 통해 휴먼 브랜드와 브랜드 진정성 및 충성도 간의 이론적인 인과관계를 규명하는 한편, 광고 선정 시 축구 스타 휴먼 브랜드의 중요성과 시사점을 도출하여 프로모션 광고 전략 수립 방안의 기초자료를 제공하고자 했다. 따라서 2018년도를 기준으로 축구 스타(선수)가 모델로 선정된 광고 영상을 시청한 경험이 있는 20세 이상 성인 을 대상으로 조사를 수행했으며, 총 382부의 자료를 최종 분석에 이용했다. 자료처리는 통계 프로그램 SPSS(ver. 21.0)과 AMOS(ver. 20.0)을 활용해 빈도분석, 신뢰도 및 타당도 검증을 위한 내적일관성과 확인적 요인분석, 변인들 간의 관계 및 방향성을 검증하는 상관분석을 실시했다. 또한 구조방정식 모형 분석(연구모형, 경로분석)을 실시했으며, 붓스트랩의 BC법을 적용해 휴먼 브랜드의 효과를 검증했다. 분석결과 첫째, 휴먼 브랜드가 브랜드 진정성에 미치는 영 향력은 통계적으로 유의하게 나타났다. 둘째, 브랜드 진정성이 충성도에 미치는 영향력은 통계적으로 유의하게 나타났 다. 셋째, 휴먼 브랜드가 충성도에 미치는 영향력은 통계적으로 유의하게 나타났다. 마지막 효과 검증 결과, 휴먼 브랜드 는 충성도에 통계적으로 유의한 간접효과 보다 직접 효과가 더 큰 것으로 나타났으며, 모두 통계적으로 유의한 것으로 나타나 매개변인 브랜드 진정성은 부분매개효과가 있는 것으로 나타났다.

The purpose of this study was to verification of human brand effect of soccer star applying bootstrap method. It also sought to provide basic information about promotion advertising strategy by human brand of soccer star. A survey was conducted on 382 adults over 20 years old who had experience watching advertisement of soccer star model by 2018. For data processing, SPSS (ver. 21.0) and AMOS (ver. 20.0) were used and frequency analysis, internal consistency, confirmatory factor analysis, correlation analysis, structural equation model analysis and BC method of bootstrap were applied. The results of this study are as follows. First, human brand of soccer star had a significant influence on brand integrity. Second, brand integrity had a significant influence on loyalty. Third, human brand of soccer star had a significant influence on loyalty. Lastly, brand integrity between human brand of soccer star and loyalty showed partial mediated effect.

시뮬레이션 출력분석 방법 중 가장 기본적이며 우수한 결과를 도출 가능한 방법은 독립반복 수행이다. 그러나 시스템이 복잡해질수록 독립 반복 시간 및 비용은 증가한다. 이러한 문제를 해결하기 위해 개발된 시뮬레 이션 출력분석방법 중 하나가 부트스트랩이다. 자기상관이 존재하는 출력 데이터 분석에 적용 가능한 대표적인 방 법으로는 이동블록 부트스트랩, 임계값 부트스트랩, 정상 부트스트랩 방법이 있다. 본 논문에서는 시스템 상태와 상태의 유지시간처럼 동시에 측정되어 이차원으로 표현되는 이변량 데이터에 대한 임계값 부트스트랩 적용 방법 을 제시한다. 특히, 두 변량의 데이터 내부에 각각 자기상관이 존재함은 물론 변량 간의 상관관계도 존재하는 경 우에 대해서 다룬다. 이 경우, 기존의 재추출 방법으로는 올바른 시스템 성능척도의 추론이 불가하다. 본 연구에서 는 이에 대한 해결책으로써 수정된 임계값 부트스트랩을 제시하며 실험을 통해 그 타당성을 검증하였다.

The independent replication is the simplest method for simulation output analysis, and facilitates correct performance evaluation. However, as the system becomes more complex, the runtime and the cost for the method increase. Bootstrap is an alternative method developed to solve the problem. There are representative methods applicable to the autocorrelated data such as moving block bootstrap, stationary bootstrap, and threshold bootstrap. In this paper, a method to apply the threshold bootstrap to bivariate data such as the system state and the state holding time is proposed. In particular, we deal with the case where autocorrelation structure exists within the data of each variate and the correlation between the two variates also exists. In this case, it is impossible to infer correct system performance measures using the existing resampling methods. This study proposes a modified threshold bootstrap as a solution of this case, and perform the coverage test to check its validity through experiments.

This paper proposes a new practical method for determining when to stop software testing. This issue has been widely known as the optimal release problem of software product, and many researchers have been developing mathematical/computational models for finding the solution for decades. We try to develop a new quality control charting to help making the right decision for it by employing the moving average model and bootstrap scheme. After discussing the modeling, we show an example of the statistical decision making of the optimal software release time.

Bootstrap method, a computer-intensive statistical technique to estimate the distribution of a statistic was appliedto deal with uncertainty and variability of the experimental data in stochastic prediction modeling of microbial growth on achill-stored food. Three different bootstrapping methods for the curve-fitting to the microbial count data were compared indetermining the parameters of Baranyi and Roberts growth model: nonlinear regression to static version function withresampling residuals onto all the experimental microbial count data; static version regression onto mean counts at samplingtimes; dynamic version fitting of differential equations onto the bootstrapped mean counts. All the methods outputted almostsame mean values of the parameters with difference in their distribution. Parameter search according to the dynamic form ofdifferential equations resulted in the largest distribution of the model parameters but produced the confidence interval of thepredicted microbial count close to those of nonlinear regression of static equation.

In this paper, long term trend data from condition monitoring systems of hydropower units was used to separate to grid cells based on the water head and active power, and to build the health sample model for abnormal evaluation. For the problem of insufficient monitoring samples on non-rated operating conditions, Bootstrap method is introduced for sample expansion and confidence interval calculation, thereby to achieve limit optimize settings of health sample of hydropower unit condition data, which is more approximate to practical engineering than normal approximation method, and has achieved a good result.

In this paper, we consider a bootstrap method to the 'row and column effects model' (RC model) to analyze a contingency table with ordered variables. We propose a bootstrap procedure for testing of independence, equality of intervals, and goodness of fit in the RC model. A real data example is included.

The k-medians clustering method is considered to partition observations into k clusters. Consistency and advantage of bootstrap confidence sets of k optimal cluster centers are discussed. The k-medians and k-means clustering methods are compared by using actual data sets.

In social network analysis, we use various measures for networks, but we don’t give these statistics with these standard errors. Generally It is useful to have an indication of how precise measures are, particularly when making inference for the network statistics. In order to calculate these standard errors, we use the bootstrap method. The basic idea of the bootstrap is that the observed data are treated as a population in itself, and that random samples of size N are drawn with replacement from the observed data. In this paper we introduce a bootstrap method for social network analysis by ties resampling and give bootstrap confidence intervals for centrality measures and density of real data, Dongeui friendship data. As a result, we found that the network density of Dongeui friendship data is 0.021, 95% bootstrap confidence intervals are 0.0174 and 0.0255. In the Dongeui friendship network, we performed a test of hypothesis that two major actors have same centrality measures, and we knew that these centrality scores are not significantly different from each other.

A new bootstrap method combined with the stationary bootstrap of Politis and Romano (1994) and the classical residual-based bootstrap is applied to stationary autoregressive (AR) time series models. A stationary bootstrap procedure is implemented for the ordinary least squares estimator (OLSE), along with classical bootstrap residuals for estimated errors, and its large sample validity is proved. A finite sample study numerically compares the proposed bootstrap estimator with the estimator based on the classical residual-based bootstrapping. The study shows that the proposed bootstrapping is more effective in estimating the AR coefficients than the residual-based bootstrapping.

Change point testing problem is considered. Kernel density estimators are used for constructing proposed change point test statistics. The proposed method can be used to the hypothesis testing of not only parameter change but also distributional change. Bootstrap method is applied to get the sampling distribution of proposed test statistic. Small sample Monte Carlo Simulation were also conducted in order to show the performance of proposed method.

Kriging is a linear interpolation method that predict values at unknown locations by using observations at known locations. The kriging variance can be used as a performance measure of a kriging model since it qualifies the accuracy of the prediction. However the classical kriging variance formula underestimates the true kriging variance. Therefore, in this paper, we propose a bootstrap kriging variance using a spatial bootstrap method and present simulation results of some models.

The estimation of 90% parametric confidence intervals (CIs) of mean AUC and Cmax ratios in bioequivalence (BE) tests are based upon the assumption that formulation effects in log-transformed data are normally distributed. To compare the parametric CIs with those obtained from nonparametric methods we performed repeated estimation of bootstrap-resampled datasets. The AUC and Cmax values from 3 archived datasets were used. BE tests on 1,000 resampled data sets from each archived dataset were performed using SAS (Enterprise Guide Ver.3). Bootstrap nonparametric 90% CIs of formulation effects were then compared with the parametric 90% CIs of the original datasets. The 90% CIs of formulation effects estimated from the 3 archived datasets were slightly different from nonparametric 90% CIs obtained from BE tests on resampled datasets. Histograms and density curves of formulation effects obtained from resampled datasets were similar to those of normal distribution. However, in 2 of 3 resampled log (AUC) datasets, the estimates of formulation effects did not follow the Gaussian distribution. Bias-corrected and accelerated (BCa) CIs, one of the nonparametric CIs of formulation effects, shifted outside the parametric 90% CIs of the archived datasets in these 2 non-normally distributed resampled log (AUC) datasets. Currently, the 80~125% rule based upon the parametric 90% CIs is widely accepted under the assumption of normally distributed formulation effects in log-transformed data. However, nonparametric CIs may be a better choice when data do not follow this assumption.

This research considers to propose the control charts using median for the location parameter. In order to decide the control limits, we apply several bootstrap methods through the approach obtaining the confidence interval except the standard bootstrap method. Then we illustrate our procedure using an example and compare the performance among the various bootstrap methods by obtaining the length between control limits through the simulation study. The standard bootstrap may be apt to yield shortest length while the bootstrap-t method, the longest one. Finally we comment briefly about some specific features as concluding remarks.

In recent years, kernel type estimates are abundant. In this paper, we propose a bandwidth selection method for kernel regression of fixed design based on bootstrap procedure. Mathematical properties of proposed bootstrap-based bandwidth selection method are discussed. Performance of the proposed method for small sample case is compared with that of cross-validation method via a simulation study.

Permutation test has been applied for the QTL(quantitative trait loci) analysis and we selected a major locus. K -means clustering analysis, for the major DNA Marker mining of ILSTS035 microsatellite loci in Hanwoo chromosome 6, has been described. Finally, bootstrap testing method has been adapted to calculate confidence intervals and for finding major DNA Markers.