• Journal of the Korean Data and Information Science Society
• /
• v.20 no.2
• /
• pp.261-272
• /
• 2009

Kolmogorov-Smirnov (K-S) statistic has been widely used for testing homogeneity of two distributions in the credit rating models. Joseph (2005) used K-S statistic to obtain validation criteria which is most well-known. There are other homogeneity test statistics such as the Cramer-von Mises, Anderson-Darling, and Watson statistics. In this paper, these statistics are introduced and applied to obtain criterion of these statistics by extending Joseph (2005)'s work. Another set of alternative criterion is suggested according to various sample sizes, type a error rates, and the ratios of bads and goods by using the simulated data under the similar situation as real credit rating data. We compare and explore among Joseph's criteria and two sets of the proposed criterion and discuss their applications.

• The Korean Journal of Applied Statistics
• /
• v.23 no.6
• /
• pp.1103-1113
• /
• 2010

An extensively used approach for family based association test(FBAT) is compared with the sib transmission/disequilibrium test(S-TDT), and in particular the adjusted S-TDT, in which the covariance among related siblings is taken into consideration, can provide a more sensitive test statistic for association. A simulation study comparing the three test statistics demonstrates that the type I error rates of all three tests are larger than the prespecified significance level and the power of the FBAT is lower than those of the other two tests. More detailed studies are required in order to assess the influence of the assumed conditions in FBAT on the efficiency of the test.

• The Korean Journal of Applied Statistics
• /
• v.21 no.6
• /
• pp.1065-1075
• /
• 2008

For the model validation of credit rating models, Kolmogorov-Smirnov(K-S) statistic has been widely used as a testing method of discriminatory power from the probabilities of default for default and non-default. For the credit rating works, K-S statistics are to test two identical distribution functions which are partitioned from a distribution. In this paper under the assumption that the distribution is known, modified K-S statistic which is formulated by using known distributions is proposed and compared K-S statistic.

• The Korean Journal of Applied Statistics
• /
• v.24 no.1
• /
• pp.35-43
• /
• 2011

In this paper, we introduce a unit root test via discrete cosine transform in the AR(1) process. We first investigate the statistical properties of DCT coefficients under the stationary AR(1) process and the random walk process in order to verify the validity of the proposed method. A bootstrapping approach is proposed to induce the distribution of the test statistic under the unit root. We performed simulation studies for comparing the powers of the Dickey-Fuller test and the proposed test.

• The Korean Journal of Applied Statistics
• /
• v.20 no.2
• /
• pp.323-332
• /
• 2007

We present a new type of portmanteau test in the frequency domain which is derived from the discrete cosine transform(DCT). For the stationary time series, DCT coefficients are asymptotically independent and their variances are expressed by linear combinations of autocovariances. The covariance matrix of DCT coefficients for white noises is diagonal matrix whose diagonal elements is the variance of time series. A simple way to test the independence of time series is that we divide DCT coefficients into two or three parts and then compare sample variances. We also do this by testing the slope in the linear regression model of which the response variables are absolute values or squares of coefficients. Simulation results show that the proposed tests has much higher powers than Ljung-Box test in most cases of our experiments.

• Journal of the Korean Data and Information Science Society
• /
• v.10 no.2
• /
• pp.387-395
• /
• 1999

The goal of this paper is to study of the entropy - based on goodness-of-fit tests with several parametric models. It is also to study the relationship of the Moran's test statistic on the view of entropy.

• Journal of the Korean Data and Information Science Society
• /
• v.24 no.2
• /
• pp.289-301
• /
• 2013

We analyze respondents' reaction to Obama's advertisement, titled 'Fix the Economy'. These respondents are divided into three groups of democratic party, republican party and independent group. By manipulating the skin complexion of the Obama photo, participants were either exposed to the dark or light version of the Obama photograph. In order to obtain decorrelated stationary data, we have applied the discrete Fourier transform to each curve and then we have applied Fan (1998)'s adaptive Neyman test to the discrete Fourier transformed data. As a result, a significant difference is found out only in the independent group.

• Journal of the Korean Data and Information Science Society
• /
• v.27 no.2
• /
• pp.373-380
• /
• 2016

Even though Likert type data is ordinal scale, many researchers who regard Likert type data as interval scale adapt as parametric methods. In this research, simulations have been used to find out a proper analysis of Likert type data. The locations and response distributions of five point Likert type data samples having diverse distribution have been evaluated. In estimating samples' locations, we considered parametric method and non-parametric method, which are t-test and Mann-Whitney test respectively. In addition, to test response distribution, we employed Chi-squared test and Kolmogorov-Smirnov test. In this study, we assessed the performance of the four aforementioned methods by comparing Type I error ratio and statistical power.

• The Korean Journal of Applied Statistics
• /
• v.8 no.1
• /
• pp.105-117
• /
• 1995

An asymptotically distribution-free procedure is proposed for paralleism of k regression lines against umbrella alternatives. Asymptotic properties are discussed, and comparative results relative to Kim and Lim(1994)'s tests show that our procedure is generally more powerful.

• Journal of the Korean Data and Information Science Society
• /
• v.25 no.5
• /
• pp.971-986
• /
• 2014

When thousands of hypotheses are tested simultaneously, the probability of rejecting any true hypotheses increases, and large multiplicity problems are generated. To solve these problems, researchers have proposed different approaches to multiple testing methods, considering family-wise error rate (FWER), false discovery rate (FDR) or false nondiscovery rate (FNR) as a type I error and some test statistics. In this article, we discuss Bonferroni (1960), Holm (1979), Benjamini and Hochberg (1995) and Benjamini and Yekutieli (2001) procedures based on T statistics, modified T statistics or local-pooled-error (LPE) statistics. We also consider Sun and Cai (2007) procedure based on Z statistics. These procedures are compared in the simulation and applied to Arabidopsis microarray gene expression data to identify differentially expressed genes.