• 대한수학회지
• /
• 제57권2호
• /
• pp.401-414
• /
• 2020

It is computationally expensive to compute principal components from scratch at every update or downdate when new data arrive and existing data are truncated from the data matrix frequently. To overcome this limitations, incremental principal component analysis is considered. Specifically, we present a sliding window based efficient incremental principal component computation from a covariance matrix which comprises of two procedures; simultaneous update and downdate of principal components, followed by the rank-one matrix update. Additionally we track the accurate decomposition error and the adaptive numerical rank. Experiments show that the proposed algorithm enables a faster execution speed and no-meaningful decomposition error differences compared to typical incremental principal component analysis algorithms, thereby maintaining a good approximation for the principal components.

• 한국항공운항학회:학술대회논문집
• /
• 한국항공운항학회 2006년도 추계학술대회 논문집
• /
• pp.61-65
• /
• 2006

The flight deck crew are under the operational time pressure in the cockpit. The crew tend to make errors when they face the threats of operational time pressure because they are in a rush. The flight deck crew can reduce threats and errors which existing within the airlines by using threat and error management when the crew know these threats and errors. The airlines can implement meaningful safety management system by analyzing into the useful information for to identify the hazard and manage the risk to reduce these threats and errors since aircrafts accidents can be fatal.With the threats and errors that were found regard to operational time pressures, company may implements safety change process to improve the safety systematically and the crew can manage the threats and errors more effectively.

• 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.

• Nuclear Engineering and Technology
• /
• 제37권2호
• /
• pp.159-166
• /
• 2005

Human reliability assessment (HRA) is conducted on the unspoken premise that 'human error' is a meaningful concept and that it can be associated with individual actions. The basis for this assumption it found in the origin of HRA, as a necessary extension of PSA to account for the impact of failures emanating from human actions. Although it was natural to model HRA on PSA, a large number of studies have shown that the premises are wrong, specifically that human and technological functions cannot be decomposed in the same manner. The general experience from accident studies also indicates that action failures are a function of the context, and that it is the variability of the context rather than the 'human error probability' that is the much sought for signal. Accepting this will have significant consequences for the way in which HRA, and ultimately also PSA, should be pursued.

• Communications for Statistical Applications and Methods
• /
• 제28권4호
• /
• pp.351-368
• /
• 2021

For a probabilistic model with positively skewed data, a lognormal distribution is one of the key distributions that play a critical role. Several lognormal models can be found in various areas, such as medical science, engineering, and finance. In this paper, we propose a new estimator for a lognormal mean and depict the performance of the proposed estimator in terms of the relative mean squared error (RMSE) compared with Shen's estimator (Shen et al., 2006), which is considered the best estimator among the existing methods. The proposed estimator includes a tuning parameter. By finding the optimal value of the tuning parameter, we can improve the average performance of the proposed estimator over the typical range of σ². The bias reduction of the proposed estimator tends to exceed the increased variance, and it results in a smaller RMSE than Shen's estimator. A numerical study reveals that the proposed estimator has performance comparable with Shen's estimator when σ² is small and exhibits a meaningful decrease in the RMSE under moderate and large σ² values.

• 전력전자학회논문지
• /
• 제27권4호
• /
• pp.332-338
• /
• 2022

This study proposes a battery state-of-health estimation method by applying a feature extraction technique. The technique that can improve estimation performance is the process of identifying and extracting meaningful data. To apply a data-driven-based aging state estimation method to batteries, health indicators are used as training data. However, limitations occur in extracting health indicators from charge/discharge cycles. This study proposes a deep-learning-based battery state-of-health estimation method that applies feature extraction techniques to compensate for this problem. According to the performance evaluation result of the proposed method, it has a low estimation error of 0.3887% based on an absolute error evaluation method.

• 대한의용생체공학회:의공학회지
• /
• 제30권1호
• /
• pp.10-17
• /
• 2009

Accurate electroencephalography (EEG) forward calculation is of importance for the accurate estimation of neuronal electrical sources. Conventional studies concerning the EEG forward problems have investigated various factors influencing the forward solution accuracy, e.g. tissue conductivity values in head compartments, anisotropic conductivity distribution of a head model, tessellation patterns of boundary element models, the number of elements used for boundary/finite element method (BEM/FEM), and so on. In the present paper, we investigated the influence of modeling errors in the boundary element volume conductor models upon the accuracy of the EEG forward solutions. From our simulation results, we could confirm that accurate construction of boundary element models is one of the key factors in obtaining accurate EEG forward solutions from BEM. Among three boundaries (scalp, outer skull, and inner skull boundary), the solution errors originated from the modeling error in the scalp boundary were most significant. We found that the nonuniform error distribution on the scalp surface is closely related to the electrode configuration and the error distributions on the outer and inner skull boundaries have statistically meaningful similarity to the curvature distributions of the boundary surfaces. Our simulation results also demonstrated that the accumulation of small modeling errors could lead to considerable errors in the EEG source localization. It is expected that our finding can be a useful reference in generating boundary element head models.

• 한국자동차공학회논문집
• /
• 제10권4호
• /
• pp.85-92
• /
• 2002

To investigate the safety of the in-use vehicles emission against the tail-pipe emission regulation, in-use vehicles emission trend according to vehicle mileage should be known. But it is impossible to collect all vehicles emission data In order to know that. Therefore, it is necessary to establish a statistically meaningful inference method that can be used generally to estimate in-use vehicles emissions distribution according to the vehicle mileage with relatively less in-use vehicles emission data. To do this, a linear regression model that solved the problems of data normality and common variance of error was studied. As a way that can secure the data normality, In(emission) instead of emission itself was used as a sampled data. And a reciprocal of mileage was suggested as a factor to secure common variance of error. As an example, 36 data of FTP-75 test were handled in this study. As a result, using average value and standard deviation at each mileage which were inferred from a linear regression model, probability density distribution and cumulative distribution of emissions according to the vehicle mileage were obtained and it was possible to predict the deterioration factor through full useful life mileage and also possible to decide whether those in-use vehicles will meet the tail-pipe emission regulations or not.

• 전자공학회논문지D
• /
• 제36D권8호
• /
• pp.19-26
• /
• 1999

개방단말 동축선 프로브의 제작오차가 프로브에 접촉된 매질의 복소유전율 측정에 얼마나 강한 영향을 미치는 지를 살펴보기 위한 수치계산을 수행하였다. 몇 가지 종류의 유전체를 접촉시킨 개방단말 동축선 프로브의 반사계수는 FDTD 방법을 적용해 수치계산 하였다. 이러한 계산된 반사계수들을 새로운 가상의 전송선 모델에 적용시켜 복소유전율로 확산하였는데, 본 논문에서 사용한 방법은 기존의 방법보다 물리적으로 더 의미가 있는 방법이다. 개방단말 프로브에 약간의 제작오차가 있을 때에도 복소유전율에는 큰 오차가 발생함을 알 수 있었다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제12권12호
• /
• pp.6000-6017
• /
• 2018

This paper presents a new descriptor, named Histograms of Prominent Edge Directions (HPED), for the recognition of facial expressions in a person-independent environment. In this paper, we raise the issue of sampling error in generating the code-histogram from spatial regions of the face image, as observed in the existing descriptors. HPED describes facial appearance changes based on the statistical distribution of the top two prominent edge directions (i.e., primary and secondary direction) captured over small spatial regions of the face. Compared to existing descriptors, HPED uses a smaller number of code-bins to describe the spatial regions, which helps avoid sampling error despite having fewer samples while preserving the valuable spatial information. In contrast to the existing Histogram of Oriented Gradients (HOG) that uses the histogram of the primary edge direction (i.e., gradient orientation) only, we additionally consider the histogram of the secondary edge direction, which provides more meaningful shape information related to the local texture. Experiments on popular facial expression datasets demonstrate the superior performance of the proposed HPED against existing descriptors in a person-independent environment.