• Communications for Statistical Applications and Methods
• /
• 제6권2호
• /
• pp.501-510
• /
• 1999

In this paper, we consider a comparable study on three Bayes procedures for the multivariate normal mean estimation problem. In specific we consider hierarchical Bayes empirical Bayes and robust Bayes estimators for the normal means. Then three procedures are compared in terms of the four comparison criteria(i.e. Average Relative Bias (ARB) Average Squared Relative Bias (ASRB) Average Absolute Bias(AAB) Average Squared Deviation (ASD) using the real data set.

• Communications for Statistical Applications and Methods
• /
• 제10권3호
• /
• pp.627-635
• /
• 2003

In this short communication we discuss the bias problem of C4.5 in split variable selection and suggest a method to reduce the variable selection bias among categorical predictor variables. A penalty proportional to the number of categories is applied to the splitting criterion gain of C4.5. The results of empirical comparisons show that the proposed modification of C4.5 reduces the size of classification trees.

• 한국과학교육학회지
• /
• 제20권2호
• /
• pp.214-220
• /
• 2000

고등학생 96명 (남 : 48명, 여: 48명)을 대상으로 자아 효능감과 화학 문제 해결력 사이의 차이(편향)를 조사하였다. 먼저 수리 문제와 개념 문제의 성공적인 해결에 대한 자신감을 묻는 자아 효능감 검사를 실시하였다. 이어서 자아 효능감 검사에서 제시된 것과 동일한 수리 문제 10문항과 개념 문제 10문항으로 학생들의 화학 문제 해결력을 측정하였다. 수리 문제 해결력 검사 점수가 개념 문제 해결력 검사보다 더 높았으나, 문제의 유형에 따른 자아 효능감의 차이는 없었다. 따라서 편향 점수는 수리 문제에 비하여 개념 문제에서 더 높았다. 수리 문제에서의 편향에 대한 이원 변량 분석 결과, 성과 사전 성취 수준 사이의 상호작용이 유의미하였다. 단순 효과를 분석한 결과 상위 수준 남학생의 편향 점수가 상위 수준 여학생보다 유의미하게 높았다. 대부분의 상위 수준 남학생은 자신의 능력을 과대 평가하였지만, 상위 수준 여학생은 과대 평가하는 경우보다 과소 평가하는 경우가 더 많았다.

• Communications for Statistical Applications and Methods
• /
• 제7권2호
• /
• pp.393-401
• /
• 2000

Many graphical methods have been suggested for obtaining an impression of a curvature in regression problem in which some covariates enter nonlinearly. However when true model does not belong to the class of additive models, graphical methods may contain a serious bias. A method is suggested which can avoid such bias in the fitting of nonaddive models.

• 대한산업공학회지
• /
• 제28권1호
• /
• pp.112-127
• /
• 2002

A useful pattern is a pattern that contributes much to learning. For a classification problem those patterns near the class boundary surfaces carry more information to the classifier. For a regression problem the ones near the estimated surface carry more information. In both cases, the usefulness is defined only for those patterns either without error or with negligible error. Using only the useful patterns gives several benefits. First, computational complexity in memory and time for learning is decreased. Second, overfitting is avoided even when the learner is over-sized. Third, learning results in more stable learners. In this paper, we propose a pattern 'utility index' that measures the utility of an individual pattern. The utility index is based on the bias and variance of a pattern trained by a network ensemble. In classification, the pattern with a low bias and a high variance gets a high score. In regression, on the other hand, the one with a low bias and a low variance gets a high score. Based on the distribution of the utility index, the original training set is divided into a high-score group and a low-score group. Only the high-score group is then used for training. The proposed method is tested on synthetic and real-world benchmark datasets. The proposed approach gives a better or at least similar performance.

• Communications for Statistical Applications and Methods
• /
• 제5권2호
• /
• pp.339-351
• /
• 1998

Designs for discriminating between two linear regression models are studied under $\Lambda$-type optimalities maximizing the measure for the lack of fit for the designs with fixed model inadequacy. The problem of selecting an appropriate $\Lambda$-type optimalities is shown to be closely related to the estimation method. $\Lambda$-type optimalities for the least squares and minimum bias estimation methods are considered. The minimum bias designs are suggested for the designs invariant with respect to the two estimation methods. First order minimum bias designs optimal under $\Lambda$-type optimalities are then derived. Finally for the case where the lack of fit test is significant, an approach to the construction of a second order design accommodating the optimal first order minimum bias design is illustrated.

• 한국정보시스템학회지:정보시스템연구
• /
• 제21권1호
• /
• pp.1-17
• /
• 2012

Common Method Bias(CMB) may cause the potential inflation of correlations between measures assessed via the same method. The problem of CMB has been well known in behavioral sciences because the survey method with self-reporting is vulnerable to CMB. Thus, the discussion on CMB is still ongoing in the MIS research in US. However, in Korea, the MIS research has never paid attention on the CMB problem. The purpose of this study is to examine the CMB problem in the Korean MIS research. To evaluate the effect of CMB, empirical studies on Technology Acceptance Model(TAM) are selected because (1) TAM is one of the MIS research areas studied intensively, (2) TAM is a theoretical model well supported by the existing empirical studies so that the result of this study would have a great ripple effect when the CMB problem turned out to be serious, (3) CMB is domain-specific. 47 TAM samples (out of 45 studies) from three Korean Journals were selected and the relevant data were collected such as correlation matrixes and the measures of the dependent variable. To find and evaluate the size of CMB, two analytic methods (Marker-Variable Technique and Method-Method Pair Technique) are employed. The result showed that there exists CMB in the Korean studies but the problem is not so serious to distort the empirical testing, compared with that of US studies. However, considering that CMB can contaminate the testing results, Korean MIS researchers should explicitly deal with the problem in designing empirical studies and collecting data.

• Communications for Statistical Applications and Methods
• /
• 제22권2호
• /
• pp.137-146
• /
• 2015

Singh (2013) considered the dual problem to the calibration of design weights to obtain a new generalized linear regression estimator (GREG) for the finite population total. In this work, we have made an attempt to suggest a way to use the dual calibration of the design weights in case of multi-auxiliary variables; in other words, we have made an attempt to give an answer to the concern in Remark 2 of Singh (2013) work. The same idea is also used to generalize the GREG estimator proposed by Deville and S$\ddot{a}$rndal (1992). It is not an easy task to find the optimum values of the parameters appear in our approach; therefore, few suggestions are mentioned to select values for such parameters based on a random sample. Based on real data set and under simple random sampling without replacement design, our approach is compared with other approaches mentioned in this paper and for different sample sizes. Simulation results show that all estimators have negligible relative bias, and the multivariate case of Singh (2013) estimator is more efficient than other estimators.

• 전기전자학회논문지
• /
• 제20권3호
• /
• pp.251-259
• /
• 2016

본 논문에서는 순차바이어스와 스위칭 타임 튜닝기법을 통한 반도체 송수신모듈(TRM : Transmitter and Receiver Module)의 강건성 향상 방법에 대해 기술한다. 기존의 회로설계는 TRM의 소형화로 인한 송신출력신호가 수신기로 유기되어 최소수신감도(MDS : Minimum Detection Signal) 개선에 초점을 맞추어졌으나, 평균고장시간(MTBF : Mean Time Between Failure)을 만족하지 못하고 빈번히 고장이 발생하는 문제가 있었다. 본 연구는 이러한 현상을 개선하는 방법으로 순차바이어스 및 스위칭 타임 튜닝기법을 제안한다. 첫 번째로 주요 고장증상 수집 및 원인을 추론하였으며, 두 번째로 개선방법을 도출하고 시스템에 적용하여 효과를 검증하였다. 제안한 방법을 적용하여 격리도 부족에 따른 빈번한 고장증상이 해소되었다.

• 한국우주과학회:학술대회논문집(한국우주과학회보)
• /
• 한국우주과학회 2010년도 한국우주과학회보 제19권1호
• /
• pp.27.5-28
• /
• 2010

The unexpected observation condition or insufficient measurement modeling can lead to uncertain measurement errors. The uncertain measurement error of orbit determination problem typically consists of noise, bias and drift. It must be removed by using a proper estimation process for better orbit accuracy. The estimation of noise and drift is not easy because of their random or unpredictable variation. On the other hand, bias is a constant difference between the mean of the measured values and the true value, so it can be simply removed. In this study, precision orbit determination with SLR observations considering range bias estimation is presented. The Yonsei Laser-ranging Precision Orbit Determination System (YLPODS) and SLR NP (Normal Point) observations of CHAMP satellite are used for this work. The SLR residual test is performed to estimate the range bias of each arc. The result shows that we can get better orbit accuracy through range bias estimation.