• Communications for Statistical Applications and Methods
• /
• v.11 no.3
• /
• pp.575-582
• /
• 2004

We explore the structure and usefulness of CERES plot as a basic tool for dealing with curvature as a function of the new predictor in generalized linear models. If a predictor has a nonlinear effect and there are nonlinear relationships among the predictors, the partial residual plot and augmented partial residual plot are not able to display the correct functional form of the predictor. Unlike these plots, the CERES plot can show the correct form. This is illustrated by simulated data.

• Steel and Composite Structures
• /
• v.25 no.3
• /
• pp.287-298
• /
• 2017

The present study aims to determine the structural response of full scaled rectangular columns under both of vertical and lateral loads using numerical methods. In the study, the composite columns considering full concrete filled circular steel tube (FCFRST) and concrete filled double-skin rectangular steel tube (CFDSRST) section types are numerically modelled using ANSYS software. Vertical and lateral loads are applied to models to assess the structural response of the composite elements. Also similar investigations are done for reinforced concrete rectangular (RCR) columns to compare the results with those of composite elements. The analyses of the systems are statically performed for both linear and nonlinear materials. In linear static analyses, both of vertical and lateral loads are applied to models as only one step. However in nonlinear analyses, while vertical loads are applied to model as only one step, lateral loads are applied to systems as step by step. The displacement and stress changes in some critical nodes and sections and contour diagrams are reported by graphs and figures. At the end of the study, it is demonstrated that the nonlinear models reveal more accurate result then those of linear models. Also, it is highlighted that composite columns provide more and more safety, ductility compared to reinforced concrete column.

• Computers and Concrete
• /
• v.20 no.1
• /
• pp.11-22
• /
• 2017

There are two methods to model the plastification of members comprising lumped and distributed plasticity. When a reinforced concrete member experiences inelastic deformations, cracks tend to spread from the joint interface resulting in a curvature distribution; therefore, the lumped plasticity methods assuming plasticity is concentrated at a zero-length plastic hinge section at the ends of the elements, cannot model the actual behavior of reinforced concrete members. Some spread plasticity models including uniform, linear and recently power have been developed to take extended inelastic zone into account. In the aforementioned models, the extended inelastic zones in proximity of critical sections assumed close to connections are considered. Although the mentioned assumption is proper for the buildings simply imposed lateral loads, it is not appropriate for the gravity load effects. The gravity load effects can influence the inelastic zones in structural elements; therefore, the plasticity models presenting the flexibility distribution along the member merely based on lateral loads apart from the gravity load effects can bring about incorrect stiffness matrix for structure. In this study, the linear flexibility distribution model is improved to account for the distributed plasticity of members subjected to both gravity and lateral load effects. To do so, a new model in which, each member is taken as one structural element into account is proposed. Some numerical examples from previous studies are assessed and outcomes confirm the accuracy of proposed model. Also comparing the results of the proposed model with other spread plasticity models illustrates glaring error produced due to neglecting the gravity load effects.

• The Korean Journal of Applied Statistics
• /
• v.15 no.1
• /
• pp.73-84
• /
• 2002

Linear regression models with inequality constraints on the coefficients are frequently used in economic models due to sign or order constraints on the coefficients. In this paper, we propose a Bayesian approach to selecting significant explanatory variables in linear regression models with inequality constraints on the coefficients. Bayesian variable selection requires computation of posterior probability of each candidate model. We propose a method which computes all the necessary posterior model probabilities simultaneously. In specific, we obtain posterior samples form the most general model via Gibbs sampling algorithm (Gelfand and Smith, 1990) and compute the posterior probabilities by using the samples. A real example is given to illustrate the method.

• Journal of the Korean Statistical Society
• /
• v.28 no.2
• /
• pp.183-193
• /
• 1999

In the problem of estimating estimable functions in classification linear models, we propose a method of obtaining least squares estimators of estimable functions. This method is based on the hierarchical Bayesian approach for estimating a vector of unknown parameters. Also, we verify that estimators obtained by our method are identical to least squares estimators of estimable functions obtained by using either generalized inverses or full rank reparametrization of the models. Some examples are given which illustrate our results.

• Journal of Korean Society of Industrial and Systems Engineering
• /
• v.23 no.55
• /
• pp.33-42
• /
• 2000

It is difficult to estimate the effective factors upon Cumulative Trauma Disorders in real workplace because those are developed by combination of various risk factors for time. The purpose of this paper was to evaluate relative level of CTDs risk factors such as task-related factors, anthropometric factors, joint deviation factors and personal factors using fuzzy linear regression models. And the models are built corresponding to each category with the survey data from telephone operators. The coefficient of fuzzy models are described as the relative level of variable to present risk factors upon CTDs.

• 제어로봇시스템학회:학술대회논문집
• /
• 1995.10a
• /
• pp.162-165
• /
• 1995

For an accurate on-line measurement of the ship's attitude the paper develops an intelligent sensing system which uses one servo-type accelerometer and two servo-type inclinometers appropriately located on the ship. By considering the dynamics of the servo-controlled rigid pendulums of the inclinometers, linear equations for the rolling and pitching of the ship are derived separately from each other. Moreover, one accelerometer is used for extracting the heaving signal. Through the introduction of linear dynamic models and the linear observation equations for the heaving, rolling and pitching, the on-line measurement of the three signals can be reduced to the state estimation of the linear dynamic systems. A bank of Kalman filters is adaptively used to achieve the on-line accurate state estimation and to overcome changes in parameters in the linear dynamic models.

• The Korean Journal of Applied Statistics
• /
• v.23 no.6
• /
• pp.1169-1178
• /
• 2010

This article consider a performance model selection based on AIC under unbalanced deign in linear mixed effect models. Vaida and Balanchard (2005) proposed conditional AIC for model selection in linear mixed effect models when the prediction of random effects is of primary interest. Theoretical properties of cAIC and related criteria have been investigated by Liang et al. (2008) and Greven and Kneib (2010). However, all of the simulation studies were performed under a balanced design. Even though functional form of AIC remain same even under the unbalanced deign, it is worthwhile to investigate performance of AIC based model selection criteria under the unbalanced design. The simulation study in this article shows how unbalancedness affects model selection in linear mixed effect models.

• Journal of Industrial Technology
• /
• v.44 no.1
• /
• pp.1-7
• /
• 2024

Since 2015, the PM_2.5 measurement data has been publicly available nationwide in South Korea, but its use is restricted to after 2015, unlike other air pollutants. To overcome this limitation, multiple linear regression and artificial neural network models were developed to predict the daily average PM_2.5 values in South Korea before 2015. The daily data of air pollution measurement(SO₂, CO, O₃, NO₂, PM₁₀) and meteorological observation data (temperature, humidity, wind speed, atmospheric pressure, precipitation, snowfall) were used as input variables to develop regional prediction models for five regions(Seoul, Incheon, Gwangju, Daejeon, Ulsan) and a national prediction model. The models were developed and validated using the air pollution measurement data after 2015, and applied to predict PM_2.5 values before 2015. The multiple linear regression model showed R² values of 0.80 nationwide, 0.73 in Seoul, and 0.67 in Incheon, which enabled estimation of daily average PM_2.5 values before 2015. The artificial neural network model showed good prediction power with R2 values of 0.79 in Gwangju, 0.81 in Daejeon, and 0.72 in Ulsan. The regional prediction models showed good prediction power in most regions, and both the multiple linear regression and artificial neural network models showed good prediction power.

• Journal of the Korean Data and Information Science Society
• /
• v.20 no.1
• /
• pp.203-209
• /
• 2009

Frailty models are useful for the analysis of correlated and/or heterogeneous survival data. However, the inferences of fixed parameters, rather than random effects, have been mainly studied. The prediction (or estimation) of random effects is also practically useful to investigate the heterogeneity of the hospital or patient effects. In this paper we propose how to extend the prediction method for random effects in HGLMs (hierarchical generalized linear models) to log-normal semiparametric frailty models with nonparametric baseline hazard. The proposed method is demonstrated by a simulation study.