• Journal of Korean Society of Forest Science
• /
• v.104 no.1
• /
• pp.117-126
• /
• 2015

Forest statistics of local administrative districts have many social needs, nevertheless we have some difficulties for working out an accurate statistics because of insufficient data in small-area level. Thus, new small-area estimation method has to set aside additional data, decrease errors of statistics and consider the local forest characteristics at the same time. In this study, we researched the spatial divisions that can set aside additional data for statistics production and satisfy the major premise, which is "forest characteristics of spatial divisions have to be equal to that of small-area". And we compared synthetic estimation methods based on three different spatial divisions(provinces, neighbor districts and new expanded districts). New expanded districts were divided based on the criteria of climate, soil type and tree species composition that affects local forest characteristics. Small-area statistics were assessed in terms of the ability to estimate local forest characteristics and consistency within large-area statistics. As a result, new expanded districts synthetic estimation was assessed to calculate statistics that reflects local forest characteristics better than other two estimation methods. Moreover, this synthetic estimation method produced the statistics that was included within 95% confidence interval of large-area statistics and was the closer to large-area statistics than the neighbor districts synthetic estimation.

• Korean Journal of Remote Sensing
• /
• v.5 no.2
• /
• pp.85-90
• /
• 1989

A new graph-theoretic clustering method based on relaxation labeling is proposed. More accurate classification is achieved by using relaxation labeling which take advantage of contextual information. Moreover, threshold selection is not needed in this method.

• The Korean Journal of Applied Statistics
• /
• v.19 no.3
• /
• pp.587-598
• /
• 2006

The number of variables exceeds the number of samples in microarray data. We propose a nonparametric local linear logistic classification procedure using orthogonal components for classifying high-dimensional microarray data. The proposed method is based on the local likelihood and can be applied to multi-class classification. We applied the local linear logistic classification method using PCA, PLS, and factor analysis components as new features to Leukemia data and colon data, and compare the performance of the proposed method with the conventional statistical classification procedures. The proposed method outperforms the conventional ones for each component, and PLS has shown best performance when it is embedded in the proposed method among the three orthogonal components.

• Communications for Statistical Applications and Methods
• /
• v.21 no.6
• /
• pp.487-500
• /
• 2014

A new method to calculate the transmittable prevalence of an epidemic disease is proposed based on a back-calculation formula. We calculated the probabilities of reactivation and of parasitemia as well as transmittable prevalence (the number of persons with parasitemia in the incubation period) of malaria in South Korea using incidence of 12 years(2001-2012). For this computation, a new probability function of transmittable condition is obtained. The probability of reactivation is estimated by the least squares method for the back-calculated longterm incubation period. The probability of parasitemia is calculated by a convolution of the survival function of the short-term incubation function and the probability of reactivation. Transmittable prevalence is computed by a convolution of the infected numbers and the probabilities of transmission. Confidence intervals are calculated using the parametric bootstrap method. The method proposed is applicable to other epidemic diseases in other countries where incidence and a long incubation period are available. We found the estimated transmittable prevalence in South Korea was concentrated in the summer with 276 cases on a peak at the $31^{st}$ week and with about a 60% reduction in the peak from the naive prevalence. The statistics of transmittable prevalence can be used for malaria prevention programs and to select blood transfusion donors.

• Journal of the Korean Statistical Society
• /
• v.20 no.1
• /
• pp.29-43
• /
• 1991

Two data-based bandwidth selectors which are optimal in the sense that they achieve n$\^$-$\frac{1}{2}$/ rate of convergence in kernel density estimation are proposed. The proposed bandwidth selectors are constructed by modifying Park and Marron's plug-in method. The first modification is taking Taylor expansion of the mean integrated squared error to two more terms than in the case of plug-in method. The second is estimating more accurately the functionals of the unknown density appeared in the minimizer of the expansion by using higher order kernels. The proposed bandwidth selectors were proved to be optimal in terms of convergence rate. According to small-sample Monte Carlo studies, the proposed bandwidth selectors showed better performance than all the other bandwidth selectors considered in the simulation.

• Communications for Statistical Applications and Methods
• /
• v.19 no.3
• /
• pp.359-365
• /
• 2012

The fuzzy set framework has a number of properties that make it suitable to formulize uncertain information in medical diagnosis. This study introduces a fuzzy diagnostic method based on the interval-valued interview chart and the interval-valued intuitionistic fuzzy weighted arithmetic average(IIFWAA) operator. An issue in the use of the IIFWAA operator is to determine the weights. In this study, we propose the occurrence information of symptoms as the weights. An illustrative example is provided to demonstrate its practicality and effectiveness.

• Communications for Statistical Applications and Methods
• /
• v.9 no.3
• /
• pp.877-887
• /
• 2002

In this paper, behaviors and properties of p-values for goodness-of-fit test are investigated. With some findings on the p-values, we consider some applications to determine sample size of a survey research using the regression equation based on a pilot study data. Regression equations are obtained by the well-known least squared method, and we find that regression lines could be formulated with only two data points, alternatively. For further studies, this works might be extended to t distributions for testing hypotheses about population mean in order to determine sample size of a prospective study. Also similar arguments could be explored for F test statistics.

• Journal of Korean Society for Quality Management
• /
• v.24 no.4
• /
• pp.1-13
• /
• 1996

The sequential sampling inspection method is an extension of the double-sampling and multiple-sampling methods and its theory is based on the sequential probability ratio test(SPRT). In this paper, the characteristics of SPRT for testing the propotion of defectives are approximated by using the estimated excess over the boundaries. The use of the estimated excess shows good performances in estimating the operating characteristic function and the average sample number of SPRT compared to the method by neglecting the excess. It also makes it possible to determine the boundary values which satisfy the desired error probabilities.

• The Korean Journal of Applied Statistics
• /
• v.21 no.3
• /
• pp.523-535
• /
• 2008

Initial work of microarray data analysis focused on identification of differentially expressed genes, and recently, the focus has moved to discovering significant sets of functionally related genes. We describe some problems of GSEA and PAGE, and propose a modified method to identify significant gene sets. The results based on a simulated experiment and real data analysis using a set of publicly available data show the superiority of the newly proposed method, GSA-AT, in detecting significant pathways with the accurate prediction.

• Journal of the Korean Data and Information Science Society
• /
• v.15 no.4
• /
• pp.717-729
• /
• 2004

The contour method which was originally designed for $2{\times}2{\times}2$ contingency table is studied for $2{\times}2{\times}K$ and $2{\times}3{\times}K$ tables. Whereas a contour plot for a $2{\times}2{\times}K$ table is represented on unit squared two dimensional plane, a contour plot of a $2{\times}3{\times}K$ table can be expressed with a regular hexahedron on three dimensional space. Based on contour plots for categorical data fitted to all possible three dimensional log-linear models, one might identify whether $2{\times}2{\times}k$ or $2{\times}3{\times}K$ tables are collapsible over the third variable.