• Journal of the Korean Data and Information Science Society
• /
• v.28 no.5
• /
• pp.1179-1189
• /
• 2017

In this study, we propose an alternative method for discriminant analysis using a multivariate empirical distribution function to express multivariate data as a simple one-dimensional statistic. This method turns to be the estimation process of the optimal threshold based on classification accuracy measures and an empirical distribution function of data composed of classes. This can also be visually represented on a two-dimensional plane and discussed with some measures in ROC curves, surfaces, and manifolds. In order to explore the usefulness of this method for discriminant analysis in the study, we conducted comparisons between the proposed method and the existing methods through simulations and illustrative examples. It is found that the proposed method may have better performances for some cases.

• Communications of the Korean Mathematical Society
• /
• v.19 no.1
• /
• pp.169-177
• /
• 2004

For a stationary field $\{X_{\b{j}},\b{j}{\;}\in{\;}{\mathbb{Z}}^d_{+}\}$ of associated random variables with distribution function $F(x)\;=\;P(X_{\b{1}}\;{\leq}\;x)$ we study strong consistency and asymptotic normality of the empirical distribution function, which is proposed as an estimator for F(x). We also consider strong consistency and asymptotic normality of the empirical survival function by applying these results.

• The Korean Journal of Applied Statistics
• /
• v.21 no.6
• /
• pp.1007-1014
• /
• 2008

Usability evaluation(UE) experiments are conducted to provide UE practitioners with guidelines for better outcomes. In UE research, significant quantities of empirical results have been accumulated in the past decades. While those results have been anticipated to integrate for producing generalized guidelines, traditional meta-analysis has limitations to combine UE empirical results that often show considerable heterogeneity. In this study, a new data mining method called weighted bi-clustering(WBC) was proposed to partition heterogeneous studies into homogeneous subsets. We applied the WBC to UE empirical results and identified two homogeneous subsets, each of which can be meta-analyzed. In addition, interactions between experimental conditions and UE methods were hypothesized based on the resulting partition and some interactions were confirmed via statistical tests.

• Journal of applied mathematics & informatics
• /
• v.7 no.2
• /
• pp.467-482
• /
• 2000

Testing equality of two and k distributions has long been an interesting issue in statistical inference. To overcome the sparseness of data points in high-dimensional space and deal with the general cases, we suggest several projection pursuit type statistics. Some results on the limiting distributions of the statistics are obtained, some properties of Bootstrap approximation are investigated. Furthermore, for computational reasons an approximation for the statistics the based on Number theoretic method is applied. Several simulation experiments are performed.

• Journal of the Korean Statistical Society
• /
• v.28 no.1
• /
• pp.45-55
• /
• 1999

Knowing the number of unit roots is important in the analysis of k-dimensional multivariate autoregressive process. In this paper we suggest simple multiple unit roots test statistics based on least squares estimator for the multivariate AR(1) process in which some eigenvalues are one and the rest are less than one in magnitude. The empirical distributions are tabulated for suggested test statistics. We have small Monte-Calro studies to compare the powers of the test statistics suggested by Johansen(1988) and in this paper.

• 한국데이터정보과학회:학술대회논문집
• /
• 2005.04a
• /
• pp.145-154
• /
• 2005

There have been some controversies on the use of the coefficient of determination for linear no-intercept model. One definition of the coefficient of determination, $R^2=\sum\;{y}{^{\hat{2}}/\sum\;{y^2}$, is being widely accepted only for linear no-intercept models though Kvalseth(1985) demonstrated some possible pitfalls in using such $R^2$. Main objective of this article is to provide a cautionary notice for use of the $R^2$ by pointing out its tricky aspects by means of empirical simulations.

• Communications for Statistical Applications and Methods
• /
• v.12 no.1
• /
• pp.149-167
• /
• 2005

The minimum disparity estimators introduced by Lindsay and Basu (1994) are studied empirically. An extensive simulation in this paper provides a location estimate of the small sample and supplies empirical evidence of the estimator performance for the univariate contaminated normal model. Empirical results show that the minimum generalized negative exponential disparity estimator (MGNEDE) obtains high efficiency for small sample sizes and dominates the maximum likelihood estimator (MLE) and the minimum blended weight Hellinger distance estimator (MBWHDE) with respect to efficiency at the contaminated model.

• Communications for Statistical Applications and Methods
• /
• v.19 no.6
• /
• pp.885-898
• /
• 2012

In this paper, we consider Bayesian approaches to partially linear models, in which a regression function is represented by a semiparametric additive form of a parametric linear regression function and a nonparametric regression function. We make a comparative study on the performance of widely used Bayesian partially linear models in terms of empirical analysis. Specifically, we deal with three Bayesian methods to estimate the nonparametric regression function, one method using Fourier series representation, the other method based on Gaussian process regression approach, and the third method based on the smoothness of the function and differencing. We compare the numerical performance of three methods by the root mean squared error(RMSE). For empirical analysis, we consider synthetic data with simulation studies and real data application by fitting each of them with three Bayesian methods and comparing the RMSEs.

• Communications for Statistical Applications and Methods
• /
• v.21 no.6
• /
• pp.539-559
• /
• 2014

We derive the exact distribution of summation for random samples from uniform distribution and then compare the exact distribution with the approximated normal distribution obtained by the central limit theorem. To check the similarity between two distributions, we consider five existing normality tests based on the difference between the target normal distribution and empirical distribution: Anderson-Darling test, Kolmogorov-Smirnov test, Cramer-von Mises test, Shapiro-Wilk test and Shaprio-Francia test. For the purpose of comparison, those normality tests are applied to the simulated data. It can sometimes be difficult to derive an exact distribution. Thus, we try two different transformations to find out which transform is easier to get the exact distribution in terms of calculation complexity. We compare two transformations and comment on the advantages and disadvantages for each transformation.

• The Korean Journal of Applied Statistics
• /
• v.29 no.1
• /
• pp.267-277
• /
• 2016

This paper considers we consider the estimation of copula parameters based on residuals in stochastic regression models. We prove that a semiparametric estimator using residual empirical distributions is consistent under some conditions and apply the results to the copula-ARMA model. We provide simulation results for illustration.