• Journal of the Korean Data and Information Science Society
• /
• v.16 no.1
• /
• pp.33-40
• /
• 2005

The estimating function method was proposed by Godambe(1985) for parameter estimation under unknown distributions for errors in the models. Threshold Autoregressive Heteroscedastic (Threshold-ARCH) models have been developed by Zakoian(1994) and Li and Li(1996) for explaining the asymmetric properties in the financial time series data. In this paper, we apply the estimating function method to the Threshold-ARCH model and show that the proposed estimators perform better than the MLE under the heavy-tailed distributions.

• Journal of the Korean Data and Information Science Society
• /
• v.17 no.3
• /
• pp.785-794
• /
• 2006

Suggested is an EM algorithm for estimation in mixture of shifted Poisson distributions with known shift parameters. For this type of mixture distribution, we have to utilize values of shift parameters to determine whether each of data belongs to some component distribution. We propose a method of estimating values of component information and then follow typical EM methodology. Simulation results show that the algorithm provides reasonable performance for the distribution.

• Journal of the Korean Data and Information Science Society
• /
• v.22 no.6
• /
• pp.1241-1250
• /
• 2011

In this paper, we develop noninformative priors for the log-normal distributions when the parameter of interest is the common mean. We developed Jeffreys' prior, th reference priors and the first order matching priors. It turns out that the reference prior and Jeffreys' prior do not satisfy a first order matching criterion, and Jeffreys' pri the reference prior and the first order matching prior are different. Some simulation study is performed and a real example is given.

• Communications for Statistical Applications and Methods
• /
• v.12 no.2
• /
• pp.323-333
• /
• 2005

In this paper, we develop a MCMC method for estimating the skew normal distributions. The method utilizing the data augmentation technique gives a simple way of inferring the distribution where fully parametric frequentist approaches are not available for small to moderate sample cases. Necessary theories involved in the method and computation are provided. Two numerical examples are given to demonstrate the performance of the method.

• IE interfaces
• /
• v.17 no.1
• /
• pp.1-12
• /
• 2004

In a modern semiconductor device manufacturing industry, statistical bin limits on wafer level test bin data are used for minimizing value added to defective product as well as protecting end customers from potential quality and reliability excursion. Most wafer level test bin data show skewed distributions. By Monte Carlo simulation, this paper evaluates methods and sample size effect regarding determination of statistical bin limits. In the simulation, it is assumed that wafer level test bin data follow the Poisson distribution. Hence, typical shapes of the data distribution can be specified in terms of the distribution's parameter. This study examines three different methods; 1) percentile based methodology; 2) data transformation; and 3) Poisson model fitting. The mean square error is adopted as a performance measure for each simulation scenario. Then, a case study is presented. Results show that the percentile and transformation based methods give more stable statistical bin limits associated with the real dataset. However, with highly skewed distributions, the transformation based method should be used with caution in determining statistical bin limits. When the data are well fitted to a certain probability distribution, the model fitting approach can be used in the determination. As for the sample size effect, the mean square error seems to reduce exponentially according to the sample size.

• Geophysics and Geophysical Exploration
• /
• v.10 no.3
• /
• pp.211-218
• /
• 2007

In this study, vertical electrical sounding (VES) data and ordinary kriging are used to identify the alluvial depth of each area that Korea Resources Corporation (KORES) conducted groundwater survey at Miryang area in Gyeongsangnam-do and Pocheon area in Gyeonggi-do from 2003 to 2004. To verify the applicability of VES data to ordianry kriging, regression analysis of VES data versus drillhole data is conducted. Comparing the alluvial depth distributions using ordinary kriging with existing drillhole data, the result shows that the depth distributions are reasonably depicted along with the topography and the basin. So, the ordinary kriging of VES data is useful to identify the alluvial depth distributions.

• Journal of the Korea Society of Computer and Information
• /
• v.25 no.7
• /
• pp.175-182
• /
• 2020

Fisher's classical method for combining independent p-values from continuous distributions is widely used but it is known to be inadequate for combining p-values from discrete probability distributions. Instead, the discrete analog of Fisher's classical method is used as an alternative for combining p-values from discrete distributions. In this paper, firstly we obtain p-values from discrete probability distributions associated with multi-sample location test data (Fisher-Pitman test and Kruskall-Wallis test data) by permutation method, and secondly combine the permutaion p-values by the discrete analog of Fisher's classical method. And we finally compare the combined p-values from both the discrete analog of Fisher's classical method and Fisher's classical method.

• Wind and Structures
• /
• v.31 no.4
• /
• pp.299-309
• /
• 2020

In this study, wind data such as speeds, loads and potential of Muğla which is located in the southwest of Turkey were statistically analyzed. The wind data which consists of hourly wind speed between 2010 and 2013 years, was measured at the 10-meters height in four different ground stations (Datça, Fethiye, Marmaris, Köyceğiz). These stations are operated by The Turkish State Meteorological Service (T.S.M.S). Furthermore, wind data was analyzed by using Log-Normal and Gamma distributions, since these distributions fit better than Weibull, Normal, Exponential and Logistic distributions. Root Mean Squared Error (RMSE) and the coefficients of the goodness of fit (R²) were also determined by using statistical analysis. According to the results, extreme wind speed in the research area was 33 m/s at the Datça station. The effective wind load at this speed is 0.68 kN/㎡. The highest mean power densities for Datça, Fethiye, Marmaris and Köyceğiz were found to be 46.2, 1.6, 6.5 and 2.2 W/㎡, respectively. Also, although Log-normal distribution exhibited a good performance i.e., lower AD (Anderson - Darling statistic (AD) values) values, Gamma distribution was found more suitable in the estimation of wind speed and power of the region.

• Communications for Statistical Applications and Methods
• /
• v.19 no.6
• /
• pp.837-847
• /
• 2012

Inverse Gaussian distribution is widely used in applications to analyze and model right-skewed data. To assess the appropriateness of the distribution prior to data analysis, Mudholkar and Tian (2002) proposed an entropy-based test of fit. The test is based on the entropy power fraction(EPF) index suggested by Gokhale (1983). The simulation results report that the power of the entropy-based test is superior compared to other goodness-of-fit tests; however, this observation is based on the small-scale simulation results on the standard exponential, Weibull W(1; 2) and lognormal LN(0:5; 1) distributions. A large-scale simulation should be performed against various alternative distributions to evaluate the power of the entropy-based test; however, the use of a theoretical method is more effective to investigate the powers. In this paper, utilizing the information discrimination(ID) index defined by Ehsan et al. (1995) as a mathematical tool, we scrutinize the power of the entropy-based test. The selected alternative distributions are the gamma, Weibull and lognormal distributions, which are widely used in data analysis as an alternative to inverse Gaussian distribution. The study results are provided and an illustrative example is analyzed.

• Speech Sciences
• /
• v.9 no.1
• /
• pp.39-47
• /
• 2002

This paper proposes an effective method to remedy the acoustic modeling problem inherent in the usual log-normal Parallel Model Composition intended for achieving robust speech recognition. In particular, the Gaussian kernels under the prescribed log-normal PMC cannot sufficiently express the corrupted speech distributions. The proposed scheme corrects this deficiency by judiciously selecting the 'fairly' corrupted component and by re-estimating it as a mixture of two distributions using data-driven PMC. As a result, some components become merged while equal number of components split. The determination for splitting or merging is achieved by means of measuring the similarity of the corrupted speech model to those of the clean model and the noise model. The experimental results indicate that the suggested algorithm is effective in representing the corrupted speech distributions and attains consistent improvement over various SNR and noise cases.