• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제9권11호
• /
• pp.4502-4518
• /
• 2015

Fisher linear discriminant analysis (LDA) is one of the most popular projection techniques for feature extraction and has been widely applied in face recognition. However, it cannot be used when encountering the single sample per person problem (SSPP) because the intra-class variations cannot be evaluated. In this paper, we propose a novel method called local similarity based linear discriminant analysis (LS_LDA) to solve this problem. Motivated by the "divide-conquer" strategy, we first divide the face into local blocks, and classify each local block, and then integrate all the classification results to make final decision. To make LDA feasible for SSPP problem, we further divide each block into overlapped patches and assume that these patches are from the same class. To improve the robustness of LS_LDA to outliers, we further propose local similarity based median discriminant analysis (LS_MDA), which uses class median vector to estimate the class population mean in LDA modeling. Experimental results on three popular databases show that our methods not only generalize well SSPP problem but also have strong robustness to expression, illumination, occlusion and time variation.

• Journal of the Korean Statistical Society
• /
• 제32권3호
• /
• pp.299-309
• /
• 2003

Let ${\pi}_1,...,{\pi}_{k}$k($\geq$2) independent logistic populations such that the cumulative distribution function (cdf) of an observation from the population ${\pi}_{i}$ is $$F_{i}\;=\; {\frac{1}{1+exp{-\pi(x-{\mu}_{i})/(\sigma\sqrt{3})}}},\;$\mid$x$\mid$<\;{\infty}$$ where ${\mu}_{i}(-{\infty}\; < \; {\mu}_{i}\; <\; {\infty}$ is unknown location mean and ${\delta}^2$ is known variance, i ＝ 1,..., $textsc{k}$. Let ${\mu}_{[k]}$ be the largest of all ${\mu}$'s and the population ${\pi}_{i}$ is defined to be 'good' if ${\mu}_{i}\;{\geq}\;{\mu}_{[k]}\;-\;{\delta}_1$, where ${\delta}_1\;>\;0$, i = 1,...,$textsc{k}$. A selection procedure based on sample median is proposed to select a subset of $textsc{k}$ logistic populations which includes all the good populations with probability at least $P^{*}$(a preassigned value). Simultaneous confidence intervals for the differences of location parameters, which can be derived with the help of proposed procedures, are discussed. If a population with location parameter ${\mu}_{i}\;<\;{\mu}_{[k]}\;-\;{\delta}_2({\delta}_2\;>{\delta}_1)$, i ＝ 1,...,$textsc{k}$ is considered 'bad', a selection procedure is proposed so that the probability of either selecting a bad population or omitting a good population is at most 1­ $P^{*}$.

• 한국ITS학회 논문지
• /
• 제12권6호
• /
• pp.10-21
• /
• 2013

본 연구에서는 교통정보시스템에서 통행시간의 이상치 자료를 제거하기 위한 복합 필터링 모형을 제시하였으며, 이는 중위절대편차법과 Voting Rule을 기반으로 하는 이중화된 필터링 모형에 해당한다. 본 모형은 중위절대편차법을 이용해 표본을 정규분포화 시키기 위한 1차 필터링을 수행하며, 이후 Voting Rule을 이용해 중위절대편차법의 적용 이후에도 남아 있는 이상치 자료를 제거하는 방식에 해당한다. 이때 Voting Rule은 표본의 통행시간과 평균통행시간의 차이가 임계치를 초과하는 경우 해당 표본을 이상치로 판정하며, 다수결의 원칙을 이용하여 이상치 자료의 비율에 따라 이상치에 대한 제거 여부를 결정한다. 일반국도 3호선의 경기도 광주시 구간을 대상으로 한 사례분석을 통해 복합 필터링 모형이 이상치 표본 만을 선택적으로 제거하여 통행시간 추정의 정확도 개선에 기여할 수 있음을 확인하였다.

• Journal of the Korean Data and Information Science Society
• /
• 제27권6호
• /
• pp.1487-1498
• /
• 2016

관리도를 사용하여 공정을 관리할 때, 일반적으로 공정 모수의 정확한 값은 알 수 없기 때문에 제1국면의 표본을 통하여 이를 추정해서 사용하고 있다. 또한 추정된 공정 모수를 이용하여 관리도를 설계하는 경우 관리한계는 관리상태에서의 런길이의 평균인 ARL (average run length)이 미리 지정한 값을 만족하도록 설정하고 있다. 그러나 런길이의 분포는 일반적으로 치우쳐져 있기 때문에, 런길이의 평균 대신 중위수를 사용하는 것이 바람직할 수 있다. 이 논문에서는 제1국면에서 추정한 모수를 사용하는 경우 부그룹의 크기에 따른 $\bar{X}$ 관리도의 성능에 대해 연구하였고, 이때 공정 평균에 대한 추정량은 전체 표본평균을 사용하고 공정 표준편차에 대해서는 5가지 추정량을 사용하여 이에 대한 영향을 살펴보았다. 기존 연구와 다른 점은 여러 가지의 부그룹 크기에 대해 모수 추정의 영향을 ARL 대신 런길이의 중위수인 MRL (median run length)에 기초하여 살펴보았으며, 두 가지 방법에 대해 그 결과를 비교하였다.

• Applied Biological Chemistry
• /
• 제43권1호
• /
• pp.67-71
• /
• 2000

주요 목재해충인 흰개미에 대하여 천연산 편백이 가지는 살충활성의 크기를 정량적으로 취급하고, 수종간 및 서로 다른 함침재료 간에 발생되는 활성의 차이를 termiticidal activity value(TAV, 살충활성치)로 정의하여 해석했다. 그 결과, 흰개미에 대한 TAV는 편백>육박나무>대만편백>후피향나무의 순이었으며, 특히 편백의 활성은 다른 3 수종보다 $5{\sim}9$배 강력했다. 또 편백의 median survival dosage(MSD, 반수치사량)는 108.8 mg이었다. 편백에 의한 목분시험과 여과지 함침시험 어느 경우에도 흰개미 살충활성은 첨가 시료중량의 대수치와 그 median survival time(MST, 반수치사일수)의 대수치와의 사이에 반비례 관계가 성립했다. 또한 목분과 여과지에 대한 함침재료간 활성차는 메탄올추출물에서 작은 반면 중성부에서 3.21배로 매우 컸다. 그러나 중성부의 살충력은 원목분의 $17{\sim}53%$로서 실험 조작과정에서 $47{\sim}83%$의 활성손실이 확인되었다.

• Communications for Statistical Applications and Methods
• /
• 제11권2호
• /
• pp.399-411
• /
• 2004

A common method of handling nonresponse in sample survey is to delete the cases, which may result in a substantial loss of cases. Thus in certain situation, it is of interest to create a complete set of sample values. In this case, a popular approach is to impute the missing values in the sample by the mean or the median of responders. The difficulty with this method which just replaces each missing value with a single imputed value is that inferences based on the completed dataset underestimate the precision of the inferential procedure. Various suggestions have been made to overcome the difficulty but they might not be appropriate for public-use files where the user has only limited information for about the reasons for nonresponse. In this note, a multiple imputation method is considered to create complete dataset which might be used for all possible inferential procedures without misleading or underestimating the precision.

• 한국농공학회논문집
• /
• 제47권2호
• /
• pp.99-110
• /
• 2005

Characteristics of concentrations of total nitrogen (TN), total phosphorus (TP), and chemical oxygen demand (COD) pollutant in percolated water at four paddy field sites (Soro, Odong, Munui, and Boeun) were investigated by a nonparametric test. Percolation rate measurement and percolated water sampling were taken during irrigation periods at $5{\sim}10$ day intervals. The normality of percolation rate and pollutant concentrations were examined using histogram, boxplot, and the Kolmogorov-Smirnov (K-S) test. Pollutant concentrations in percolated water showed positively skewed distribution. The median concentrations of pollutant were 1.91 mg/L for TN, 0.021 mg/L for TP, and 6.6 mg/L for COD, which were lower than its arithmetic mean concentrations by $35\%$ for TN, $36\%$ for TP, and $13\%$ for COD. The median concentrations of TN and TP differed significantly among sample sites according to the Kruskal-Wallis test. However, median concentrations were not significantly different among month except for TN and TP of Soro and COD of Odong. The percolation load of pollutants during irrigation periods in the study area were estimated at $3.12{\sim}7.75\;kg/ha$ for TN, $0.033{\sim}0.155\;kg/ha$ for TP, and 10.7 kg/ha for COD, which were much lower than respective values reported in Japan.

• 기술사
• /
• 제21권4호
• /
• pp.19-32
• /
• 1988

Whereas is non-symmetrical distribution manufacturing process the traditional X-chart by Shewhart is not plotted relatively on the central line but plotted on the skew of upper-hand side or lower-hand side. That is to say, for the purpose of producing either upper-specification-oriented items or lower-specification-oriented items, and when we carry out tighter control so as to have them pass only its specifications, the distribution shape naturally has a non-normal distribution. In the Shewhart X-chart, which is the most widely used one in Korea, such skewed distributions make tile plots to be inclined below or above the central line or outside the control limits although no assignable causes can be found. To overcome such short comings is non-normally distributed processes, a distribution-free type of confidence interval can be used, which should be haled on order statistics. This thesis is concerned with the design of control chart based on a sample median which is easy to use in practical situation and therefore properties for non-normal distributions, such as Gamma, Beta, Lognormal, Weibull, Pareto, and Truncated-normal distributions, may be easily analyzed. To enhance this improvement, I proved the property of practical applications of control chart method by comparing and analyzing the case studies of practical application of special purpose control chart method, and also by introducing the new designed median control chart.

• 응용통계연구
• /
• 제22권2호
• /
• pp.355-364
• /
• 2009

A simple method is proposed for constructing nonparametric confidence intervals for the difference or ratio of two median failure times. The method applies when clustered survival data with censoring is randomized either (I) under cluster randomization or (II) subunit randomization. This method is simple to calculate and is based on non-parametric density estimation. The proposed method is illustrated with the otology study data and HL-A antigen study data. Moreover, the simulation results are reported for practical sample sizes.

• 한국국방경영분석학회지
• /
• 제22권2호
• /
• pp.126-138
• /
• 1996

The statistical control chart has been proven to be the effective tool most widely used in the manufacturing industry for monitoring and controlling the manufacturing processes. However, the Shewhart chart sometimes gives us false information when the distribution of quality characteristics is skewed. Therefore, it cannot serve as the universal quality control chart if there exist odd events in the manufacturing process. The objective of this study is thus to develop the new technique for constructing the limits of quality control chart based on a sample median and range when the distribution of the underlying population is skewed. This new control chart can effectively solve and manage the processes which have the non-normally distributed quality characteristics frequently occurring in the practical situation.