• Journal of the Korean Statistical Society
• /
• v.29 no.2
• /
• pp.201-217
• /
• 2000

The method of generalized estimating equations (GEE) has become very popular for the analysis of longitudinal data. We extend this work to the use of M-estimators; the resultant regression estimates are robust to heavy tailed errors and to outliers. The proposed method does not require correct specification of the dependence structure between observation, and allows for heterogeneity of the error. However, an estimate of the dependence structure may be incorporated, and if it is correct this guarantees a higher efficiency for the regression estimators. A goodness-of-fit test for checking the adequacy of the assumed M-estimation regression model is also provided. Simulation studies are conducted to show the finite-sample performance of the new methods. The proposed methods are applied to a real-life data set.

• Communications for Statistical Applications and Methods
• /
• v.18 no.6
• /
• pp.751-759
• /
• 2011

The method of generalized estimating equations(GEE) is widely used in the analysis of a correlated dataset that consists of repeatedly observed responses within subjects. The GEE uses a quasi-likelihood equations to find the parameter estimates without assuming a specific distribution for the correlated responses. In this paper we study the importance of specifying the working correlation structure appropriately in fitting GEE for correlated counts data. We investigate the empirical coverages of confidence intervals for the regression coefficients according to four kinds of working correlations where one structure should be specified by the users. The confidence intervals are computed based on the asymptotic normality and the sandwich variance estimator.

• Journal of the Korean Statistical Society
• /
• v.33 no.4
• /
• pp.367-379
• /
• 2004

This article deals with the nonparametric analysis of longitudinal data when there exist possible correlations among repeated measurements for a given subject. We consider a quasi-likelihood regression model where a transformation of the regression function through a link function is linear in time-varying coefficients. We investigate the local polynomial approach to estimate the time-varying coefficients, and derive the asymptotic distribution of the estimators in this quasi-likelihood context. A real data set is analyzed as an illustrative example.

• Communications for Statistical Applications and Methods
• /
• v.27 no.2
• /
• pp.201-210
• /
• 2020

Baseline-category logit random effects models have been used to analyze longitudinal nominal data. The models account for subject-specific variations using random effects. However, the random effects covariance matrix in the models needs to explain subject-specific variations as well as serial correlations for nominal outcomes. In order to satisfy them, the covariance matrix must be heterogeneous and high-dimensional. However, it is difficult to estimate the random effects covariance matrix due to its high dimensionality and positive-definiteness. In this paper, we exploit the modified Cholesky decomposition to estimate the high-dimensional heterogeneous random effects covariance matrix. Bayesian methodology is proposed to estimate parameters of interest. The proposed methods are illustrated with real data from the McKinney Homeless Research Project.

• Proceedings of the Korea Concrete Institute Conference
• /
• 2005.11a
• /
• pp.197-200
• /
• 2005

This paper presents an analytical model for evaluation of Tension Stiffening Effect by actual Bond-Slip relationships between the reinforcement and the surrounding concrete. The presence of longitudinal splitting cracks was found to significantly after the tension stiffening. The model is applied to the longitudinal splittings cracks and derived to Tension stiffening model. The predicted values are shown to be in good agreement with the experimentally measured data.

• Journal of The Korean Association For Science Education
• /
• v.13 no.3
• /
• pp.317-326
• /
• 1993

The purpose of this study was to examine the longitudinal achievement of Science Inquiry Skills. The study compared two data sets collected in May 1992 and April 1993 using the TSIS(Test of Science Inquiry Skills). TSIS consists of 36items which measure 12 science process skills. TSIS's reliability(K-R 20) is O. 86 and test-retest reliability is O.7822.

• Proceedings of the Korea Concrete Institute Conference
• /
• 1992.04a
• /
• pp.131-136
• /
• 1992

Eight singly reinforced high strength concrete beams without web reinforcement were tested to investigate their behavior and to determine their ultimate shear capacities. In this study, the main variable was the ration of longitudinal reinforcement. Test results were compared with strength predicted by using ACI code, Zsutty's dquation and Bazant & Kim's equation. As the result, for the beams of low steel ratio, the margin of safety in ACI code may be disappeared. It was shown that Zsutty.s equation well predict the trend of the test data.

• Journal of Korea Water Resources Association
• /
• v.48 no.4
• /
• pp.291-298
• /
• 2015

The aim of this study is that a theoretical formula for estimating the one-dimensional longitudinal dispersion coefficient is derived based on a transverse distribution equation for the depth averaged stream-wise velocity in open channel. In "Part I. Theoretical equation for stream-wise velocity" which is the former volume of this article, the velocity distribution equation is derived analytically based on the Shiono-Knight Model (SKM). And then incorporating the velocity distribution equation into a triple integral formula which was proposed by Fischer (1968), the one-dimensional longitudinal dispersion coefficient can be derived theoretically in "Part II. Longitudinal dispersion coefficient" which is the latter volume of this article. SKM has presented an analytical solution to the Navier-Stokes equation to describe the transverse variations, and originally been applied to straight and nearly straight compound channel. In order to use SKM in modeling non-prismatic and meandering channels, the shape of cross-section is regarded as a triangle in this study. The analytical solution for the velocity distribution is verified using Manning's equation and applied to velocity data measured at natural streams. Although the velocity equation developed in this study do not agree well with measured data case by case, the equation has a merit that the velocity distribution can be calculated only using geometric data including Manning's roughness coefficient without any measured velocity data.

• Journal of the Korean Geographical Society
• /
• v.48 no.2
• /
• pp.272-287
• /
• 2013

This paper describes the importance of selecting explanatory variables(e.g. land price determinants) in hedonic pricing models employed in predicting real estate price, and explores dynamics of the land price determinants over time. The City of Junju was chosen as the study area, and repeated measured price data of standard lots over 17 years were analyzed. We applied a three-level modeling approach to this data in consideration of its nested data structure and longitudinal characteristics. Main land price determinants we focused on are primarily based on items included in the standard comparison table of land price, which is an official hedonic pricing model used by Government to estimate land price for tax levy. Our result shows that the land price fluctuation over 17 years was not uniform over the whole study area with each neighborhood revealing different price trend, and as such warrants longitudinal model components. In addition, some of determinants previously recognized as important were proved insignificant. It was also found that significant determinants at a particular time point lost its power gradually over time and vice versa. It is expected that more accurate prediction of price would be possible when taken account for this dynamics of price determinants over time in applying hedonic pricing model method.

• Survey Research
• /
• v.18 no.4
• /
• pp.25-59
• /
• 2017

The current study investigated how elderly husbands and wives' marital satisfaction and depression change over time and examined how changes in marital satisfaction account for changes in depression between and within the dyads. The longitudinal dyadic data from the first wave(2006) to the fifth wave(2014) of the Korean Longitudinal Study of Aging(KLoSA), collected by the Korea Employment Information Service, were used for the analyses. Because husbands and wives are interdependent within couples, we applied statistical models for dyadic data. The dyadic growth model(DGM) was used to model the trajectories of marital satisfaction and depressive symptoms. In order to analyze the association of these growth factors, we proposed the dyadic growth actor-partner interdependence model(DG-APIM) and applied the model to the data. The results showed that on average the husbands' marital satisfaction was higher but decreased faster over the course of the study than the wives'. It also showed that the average depression of the husbands was lower than that of the wives but the husbands' depression increased faster than the wives' over the course of the study. The variance of the averages of husbands' (wives's) depression was accounted for by that of wives'(husbands') marital satisfaction, showing a partner effect. The variance of the slopes of husbands'(wives') depression was accounted for by that of marital satisfaction of themselves, showing an actor effect. The results showed that there is a longitudinal interdependence between husbands and wives' marital satisfaction and depression and supported the marital discord model of depression.