• Communications for Statistical Applications and Methods
• /
• v.26 no.3
• /
• pp.273-293
• /
• 2019

We consider the problem of model selection in multiple linear regression with outliers and non-normal error distributions. In this article, the robust model selection criterion is proposed based on the robust estimation method with the least absolute deviation (LAD). The proposed criterion is shown to be consistent. We suggest proposed criterion based algorithms that are suitable for a large number of predictors in the model. These algorithms select only relevant predictor variables with probability one for large sample sizes. An exhaustive simulation study shows that the criterion performs well. However, the proposed criterion is applied to a real data set to examine its applicability. The simulation results show the proficiency of algorithms in the presence of outliers, non-normal distribution, and multicollinearity.

• The Journal of the Korean dental association
• /
• v.16 no.12 s.115
• /
• pp.923-928
• /
• 1978

The auther studied the open-bite of Hellman dental age IIIC groups of Korean childern on the roentgenocephalometry. The subjects consisted of 45 males and 45 females with normal occlusion and acceptable profile. The major conclusions may be listed as follows: 1. This study was summerized in the form of tables, mean, standard deviation, minium and maximum for 35 angular, linear and ration measurements. 2. Charts of standard deviation from the mesurements were made for the clinical application.

• The Journal of the Society of Korean Medicine Diagnostics
• /
• v.15 no.2
• /
• pp.203-214
• /
• 2011

Objectives: The objective of the current study is to determine whether an ultrasound device system is adequate for measuring distance from the abdominal surface of Sangwan (CV13) to peritoneum in front of anterior surface of liver. Methods: We recruited 3 healthy young male subjects and 2 sonographers. The each sonographer measured vertical shortest distance from the abdominal surface of Sangwan (CV13) to peritoneum in front of anterior surface of liver with a ultrasound device with three methods of 3.5 MHz convex probe and gel, 9 MHz linear probe and gel, and 9 MHz linear probe and solid gel pad, three times in random order. Because the total variation could be divided into repeatability, reproducibility and subject-to-subject variation in Gage R&R method, we compared the sources of variation associated with the measurement system with an analysis of variance model. Results & Conclusions: Number of distinct categories is calculated on the basis of standard deviation of subject-to-subject divided by standard deviation of total Gage R&R. If the number of categories is five or more, the measurement system may be acceptable for the analysis of the process. The number of distinct categories of the ultrasound device system for measuring distance from the abdominal surface of Sangwan (CV13) to peritoneum in front of anterior surface of liver were 9.20 (9 MHz linear probe and gel), 14.70 (9 MHz linear probe and solid gel pad). So we concluded that repeatabilities and reproducibilities of the ultrasound device system for measuring distance from the abdominal surface of Sangwan (CV13) to peritoneum in front of anterior surface of liver with the methods of 9 MHz linear probe and gel, and 9 MHz linear probe and solid gel pad were acceptable.

• The Korean Journal of Applied Statistics
• /
• v.17 no.3
• /
• pp.431-443
• /
• 2004

For the purpose of disease mapping, we consider the four small area estimation techniques to estimate the mortality rate of small areas; direct, Empirical estimation with total moment estimator and local moment estimator, Estimation based on hierarchial generalized linear model. The estimators are compared by empirical study based on lung cancer mortality data from 2000 Annual Reports on the Cause of Death Statistics in Gyeongsang-Do and Jeonla-Do published by Korean National Statistical Office. Also he stability and efficiency of these estimators are investigated in terms of mean square deviation as well as variation of estimates.

• Journal of Dental Rehabilitation and Applied Science
• /
• v.27 no.2
• /
• pp.175-184
• /
• 2011

With a development of implant restoration technique, there are increasing use of computer-guided system for edentulous patients. It was carried out simulated operation based on CT information about patient's bone quantity, quality and anatomical landmark. However, there are some difference between the programmed implant and post-operative implant about it's position. If the deviation was severe, it could happen a failure of 'passive fit' and not suited for path of implant restoration. The aim of this presentation is to evaluate about a degree of deviations between programmed implant and post-operative implant. Five patients treated by 'NobelGuide' system (Nobel Biocare AB, G$\ddot{o}$teborg, Sweden) in Department of Prosthodontics, Inha University were included in this study. The patients were performed CT radiograph taking and intra-oral impression taking at pre-operation. Based on CT images and study model, surgical stent was produced by NobelBiocareTM. To fabricated a pre-operative study model, after connected lab analog to surgical template, accomplished a pre-operative model using type 4 dental stone. At final impression, a post-operative study model was fabricated in the conventional procedures. Each study model was performed CT radiograph taking. Based on CT images, each implant was simulated in three dimensional position using $Procera^{(R)}$ software (Procera Software Clinical Design Premium, version 1.5; Nobel Biocare AB). In 3D simulated model, length and angulation between each implant of both pre- and post-operative implants were measured and recorded about linear and angular deviation between pre-and post-operative implants. A total of 24 implants were included in this study and 58 inter-implant sites between each implant were measured about linear and angular deviations. In the linear deviation a mean deviation of 0.41 mm (range 0~1.7 mm) was reported. In the angular deviation, a mean deviation was $1.99^{\circ}$ (range $0^{\circ}{\sim}6.7^{\circ}$). It appears that the both linear and angular mean deviation value were well acceptable to application of computer-guided implant system.

• Journal of the Korean Data and Information Science Society
• /
• v.27 no.6
• /
• pp.1653-1660
• /
• 2016

Smoothly clipped absolute deviation (SCAD) penalty is known to satisfy the desirable properties for penalty functions like as unbiasedness, sparsity and continuity. In this paper, we deal with the regression function estimation and variable selection based on SCAD penalized censored regression model. We use the local linear approximation and the iteratively reweighted least squares algorithm to solve SCAD penalized log likelihood function. The proposed method provides an efficient method for variable selection and regression function estimation. The generalized cross validation function is presented for the model selection. Applications of the proposed method are illustrated through the simulated and a real example.

• 한국신재생에너지학회:학술대회논문집
• /
• 2008.10a
• /
• pp.262-267
• /
• 2008

The long-term wind data are reconstructed from the short-term meteorological data to design the 4 MW offshore wind park which will be constructed at Woljeong-ri, Jeju island, Korea. Using two MCP (Measure-Correlate-Predict) models, the relative deviation of wind speed and direction from two neighboring reference weather stations can be regressed at each azimuth sector. The validation of the present method is checked about linear and matrix MCP models for the sets of measured data, and the characteristic wind turbulence is estimated from the ninety-percent percentile of standard deviation in the probability distribution. Using the Gumbel's model, the extreme wind speed of fifty-year return period is predicted by the reconstructed long-term data. The predicted results of this analysis concerning turbulence intensity and extreme wind speed are used for the calculation of fatigue life and extreme load in the design procedure of wind turbine structures at offshore wind farms.

• Journal of the Korean Operations Research and Management Science Society
• /
• v.11 no.1
• /
• pp.36-43
• /
• 1986

In this study, the fuzzy programming is extended to handle various types of membership functions by transformation of the complicated fuzzy programming problems into the equivalent crisp linear programming problems with single objective. It is well-known that the fuzzy programming problem with linear membership functions (i.e., ramp type) can be easily transformed into a linear programming problem by introducing one dummy variable to minimize the worst unwanted deviation. However, until recently not many researches have been done to handle various general types of complicated linear membership functions which might be more realistic than ramp-or triangular-type functions. In order to handle these complicated membership functions, the goal dividing concept, which is based on the fuzzy set operation (i. e., intersection and union operations), has been prepared. The linear model obtained using the goal dividing concept is more efficient and single than the previous models [4, 8]. In addition, this result can be easily applied to any nonlinear membership functions by piecewise approximation since the membership function is continuous and monotone increasing or decreasing.

• 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.

• Journal of Korean Society of Transportation
• /
• v.27 no.6
• /
• pp.119-127
• /
• 2009

It has been known that bus route and company related characteristics have influences on punctuality, but fewer research have been conducted. Independent variables used in this study were selected using correlation analysis, and OLS(Ordinary Least Square) and HLM(Hierarchical Linear Model) were employed to identify factors affecting bus punctuality(headway deviation). The results showed that ICC(intraclass Correlation Coefficient) is 0.10, indicating that hierarchical linear models are more adequate for these data because there is effective variation in the subjects between companies. Punctuality was found to be negatively associated with the number of vehicles, the number of persons per vehicle, and total travel time. On the other hand, average headway and company size have a positive relationship with punctuality. Therefore, the number of vehicles per route, average headway, and the number of vehicles managed by a company should be considered for more accurately evaluating the management of piunctuality.