Comparisons Between Model Selection Criteria

Choongrak Kim,Hyoungsoon Kim,Meeseon Jeong 한국통계학회 1997 Communications for statistical applications and me Vol.4 No.1

Comparison of breast volume change between oncoplastic breast-conserving surgery with radiation therapy and a simultaneous contralateral balancing procedure through the inverted-T scar technique

Kim, Min Wook,Oh, Won Seok,Lee, Jae Woo,Kim, Hyun Yul,Jung, Youn Joo,Choo, Ki Seok,Nam, Kyung Jin,Bae, Seong Hwan,Kim, Choongrak,Nam, Su Bong,Joo, Ji Hyeon Korean Society of Plastic and Reconstructive Surge 2020 Archives of Plastic Surgery Vol.47 No.6

Background Reduction mammoplasty or mastopexy is performed as an additional balancing procedure in patients with large or ptotic breasts who undergo breast-conserving surgery (BCS). Radiation therapy on breasts that have undergone surgery may result in changes in the volume. This study presents a comparative analysis of patients who received post-BCS balancing procedures to determine whether volume changes were larger in breasts that received radiation therapy than on the contralateral side. Methods Thirty-six participants were selected among patients who received BCS using the inverted-T scar technique between September 2012 and July 2017, were followed up for 2 or more years, and had pre-radiation therapy computed tomography images and post-radiation therapy images taken between 12 and 18 months after completion. The average age of the participants was 53.5 years, their average body mass index was 26.62 kg/㎡. Results The pre- and post-radiation therapy volumes of the breasts receiving BCS were 666.08±147.48 mL and 649.33±130.35 mL, respectively. In the contralateral breasts, the volume before radiation therapy was 637.69±145.72 mL, which decreased to 628.14±166.41 mL after therapy. The volume ratio of the affected to the contralateral breasts was 1.05±0.10 before radiation therapy and 1.06±0.12 after radiation therapy. Conclusions The ratio of the volume between the two breasts immediately after surgery and at roughly 18 months postoperatively was not significantly different (P=0.98). For these reasons, we recommend a simultaneous single-stage balancing procedure as a reasonable option for patients who require radiation therapy after BCS without concerns regarding volume change.

On the Selection of Bezier Points in Bezier Curve Smoothing

Kim, Choongrak,Park, Jin-Hee The Korean Statistical Society 2012 응용통계연구 Vol.25 No.6

Nonparametric methods are often used as an alternative to parametric methods to estimate density function and regression function. In this paper we consider improved methods to select the Bezier points in Bezier curve smoothing that is shown to have the same asymptotic properties as the kernel methods. We show that the proposed methods are better than the existing methods through numerical studies.

Quasi-Likelihood Regression for Varying Coefficient Models with Longitudinal Data

Choongrak Kim,Meeseon Jeong,Woochul Kim,Byeong U. Park 한국통계학회 2004 Journal of the Korean Statistical Society Vol.33 No.4

This articledeals with the nonparametricanalysis oflongitudinal datawhen there exist possible correlationsamong repeated measurements fora given subject. We consider a quasi-likelihood regressionmodel where atransformationoftheregressionfunctionthrough a link functionislinearintime-varying coe cients.We investigatethe local polynomial approach toestimatethetime-varyingcoe cients,and derivetheasymptoticdistributionoftheestimatorsin thisquasi-likelihood context.A realdata setisanalyzedas an illustrative example.

Testing the Goodness of Fit of a Parametric Model via Smoothing Parameter Estimate

Kim, Choongrak The Korean Statistical Society 2001 Journal of the Korean Statistical Society Vol.30 No.4

In this paper we propose a goodness-of-fit test statistic for testing the (null) parametric model versus the (alternative) nonparametric model. Most of existing nonparametric test statistics are based on the residuals which are obtained by regressing the data to a parametric model. Our test is based on the bootstrap estimator of the probability that the smoothing parameter estimator is infinite when fitting residuals to cubic smoothing spline. Power performance of this test is investigated and is compared with many other tests. Illustrative examples based on real data sets are given.

녹내장성 시야결손의 점별 모델링

김재현(Jaehyun Kim),문상우(Sang Woo Moon),김충락(Choongrak Kim),이지웅(Ji Woong Lee) 대한안과학회 2022 대한안과학회지 Vol.63 No.11

목적: 한국인 녹내장 환자의 시야검사 각 지점의 변화를 분석하여 회귀 모델의 적합도와 예측 성능을 비교하였다. 대상과 방법: 6년 이상 시행한 시야검사가 10회 이상인 개방각녹내장과 녹내장의증 환자에서 시야검사 각 지점에서 시간에 따른 역치값 변화에 선형, 지수형, 다항형 회귀분석을 시행하였다. 평균제곱근오차(root mean squared error, RMSE)를 이용하여, 세 모델의적합도를 비교하였고, 첫 5년의 시야검사로 모델 추정 후, 1, 2, 3, 5년 후의 역치값을 예측하여 예측 성능을 비교하였다. 중심부와주변부 영역 및 초기 역치값에 따라 예측 성능을 비교하였다. 결과: 280명 439안의 평균 경과 관찰 기간과 시야검사 개수는 9.64년과 13.02개였다. 적합도 RMSE는 다항형이 가장 낮았다(p<0.001). 1년 예측에서 선형이, 3, 5년 예측에서 지수형의 RMSE가 가장 낮았다(모든 p<0.001). 1, 2년 예측 시, 초기 역치값 0-7 decibel (dB)에서 지수형이, 20-27 dB에서 선형이 RMSE가 가장 낮았다(모든 p<0.001). 선형과 지수형 모델을 비교하였을 때, 시야 주변부 1년 예측에서 선형이, 3, 5년 예측에서 지수형이 RMSE가 더 낮았다(모든 p<0.001). 시야 중심부 2년 예측부터 지수형의 RMSE가 선형보다더 낮았다(모든 p≤0.015). 결론: 한국인 녹내장성 시야결손의 점별 분석에서 단기 예측 시에 선형이, 장기 예측 시에 지수형이 성능이 우수하였다. 중심부 시야와 초기 역치값이 낮은 경우에는 지수형의 예측 성능이 선형보다 우수하였다. Purpose: To evaluate the utility of pointwise modeling for predicting visual field (VF) progression in Korean glaucoma patients. Methods: Open-angle glaucoma or glaucoma suspect patients with VFs ≥ 10 times, who were followed-up for ≥ 6 years, were included. Linear, exponential, and polynomial regression of threshold values at each test point against time were performed. Model fit was evaluated based on root mean squared error (RMSE) for the entire longitudinal VF series. To evaluate prediction ability, VFs from the first 5 years were used to estimate model parameters, followed by calculation of threshold values for 1, 2, 3, 5 years to obtain RMSE. Prediction ability was compared regarding initial threshold value and also central and peripheral VF area. Results: Four hundred thirty-nine eyes (280 patients) were included. The mean follow-up duration and number of VF tests were 9.64 years and 13.02, respectively. When fitting the entire VF series, polynomial model had the lowest RMSE (p < 0.001). For 1-year predictions, linear model had the lowest RMSE, while exponential model had the lowest RMSE for 3- and 5-year predictions (p < 0.001). For 1- and 2-year predictions, exponential and linear models had the lowest RMSEs, with initial sensitivities of 0-7 and 20-27 decibel (dB), respectively (p < 0.001). Compared to exponential model, linear model had lower RMSE for 1-year, but higher RMSE for 3- and 5-year at peripheral VF area (p < 0.001). For central VF area, exponential model had lower RMSEs for 2-, 3-, and 5-year predictions compared to linear model (p ≤ 0.015). Conclusions: The linear model outperformed the exponential model for short-term predictions, while the exponential model was better for long-term predictions. The prediction performance of the exponential model was superior to that of the linear model for central VFs, and for test points with lower initial sensitivities.

Times to Discontinue Antidepressants Over 6 Months in Patients with Major Depressive Disorder

WooYoung Jung,SaeHeon Jang,SungGon Kim,YoungMyo Jae,BoGeum Kong,HoChan Kim,ByeongMoo Choe,JeongGee Kim,ChoongRak Kim 대한신경정신의학회 2016 PSYCHIATRY INVESTIGATION Vol.13 No.4

Objective-The aim of the present study was to investigate differences in discontinuation time among antidepressants and total antidepressant discontinuation rate of patients with depression over a 6 month period in a naturalistic treatment setting. Methods-We reviewed the medical records of 900 patients with major depressive disorder who were initially prescribed only one kind of antidepressant. The prescribed antidepressants and the reasons for discontinuation were surveyed at baseline and every 4 weeks during the 24 week study. We investigated the discontinuation rate and the mean time to discontinuation among six antidepressants groups. Results-Mean and median overall discontinuation times were 13.8 and 12 weeks, respectively. Sertraline and escitalopram had longer discontinuation times than that of fluoxetine, and patients who used sertraline discontinued use significantly later than those taking mirtazapine. No differences in discontinuation rate were observed after 24 weeks among these antidepressants. About 73% of patients discontinued antidepressant treatment after 24 weeks. Conclusion-Sertraline and escitalopram tended to have longer mean times to discontinuation, although no difference in discontinuation rate was detected between antidepressants after 24 weeks. About three-quarters of patients discontinued antidepressant maintenance therapy after 24 weeks.

Genetic structure and variability of the working dog inferred from microsatellite marker analysis

Kwon, Yun-Jeong,Choi, Bong-Hwan,Eo, Jungwoo,Kim, Choongrak,Jung, Yi-Deun,Lee, Ja-Rang,Choi, Yuri,Gim, Jeong-An,Lee, Dong-Hoon,Ha, Ji-Hong,Kim, Dae-Soo,Huh, Jae-Won,Kim, Tae-Hun,Seong, Hwan-Hoo,Kim, He Springer-Verlag 2014 Genes & Genomics Vol.36 No.2

Working dogs serve as military watch dogs, search dogs, rescue dogs, and guide dogs with un-come-at-able character. They are drafted by in-training examination including concentration, capacity for locomotion, boldness and earthly desires. In this study, genetic diversity and relationships among two groups of working dogs (pass and fail group in-training examination) were assessed based on 15 microsatellite markers in 25 individuals of working dogs (military watch dogs and Korean search dogs). For the 15 microsatellite markers, the values of allelic richness (A (R) ) ranged from 2.21 (pass group) to 1.60 (fail group) in military watch dogs, while A (R) ranged from 2.79 (pass group) to 2.72 (fail group) in Korean search dogs. Among 52 different alleles of military watch dogs, 22 alleles were detected in pass group only, while 8 alleles in fail group only. In case of Korean search dogs, 3 alleles were observed in pass group only, while 13 alleles in fail group only. These group-specific unique alleles reflect good biomarker for selecting working dogs (military watch dogs and Korean search dogs), indicating that those group specific microsatellite alleles could separate working dogs to be pass or fail group in out-training dog population. Taken together, this study demonstrates the feasibility of microsatellite analyses for the selection of superior working dogs objectively. Furthermore, this approach could be used for the proper selection of working dogs in combination with in-training examination.

Attachment distribution of pectoral muscle origins identified in dual-plane breast implant insertion

Nam, Su Bong,Song, Kyung Ho,Seo, Jung Yeol,Choi, June Seok,Park, Tae Seo,Lee, Jae Woo,Kim, Ju Hyung,Kim, Min Wook,Kim, Hyun Yeol,Jung, Yun Ju,Kim, Choongrak Korean Society of Plastic and Reconstructive Surge 2020 Archives of Plastic Surgery Vol.47 No.2

Background Implant-based dual-plane augmentation mammoplasty requires accurate separation of the pectoralis major muscle (PMM) at its origins. The authors identified the PMM origins during breast reconstruction surgery with the goal of providing additional information on subpectoral implant insertion for reconstructive or aesthetic purposes. Methods This study was conducted on 67 patients who underwent breast reconstruction surgery at the breast center of our hospital between November 2016 and June 2018. In total, 34 left and 39 right hemithoraces were examined. The left and right hemithoraces were each divided into 15 zones to determine the percentage of PMM attachments in each zone. The distribution of PMM origins in each zone was examined to identify any statistically significant differences. Results There were no statistically significant differences in the origins of the PMM between the right and left hemithoraces. The percentage of attachments increased moving from the fourth to the sixth rib and from the lateral to the medial aspect. Conclusions The anatomical findings of this study could be used as a reference for accurate dissection of the origins of the PMM for the preparation of the subpectoral pocket for subpectoral implant placement.

Nonparametric Estimation of Distribution Function using Bezier Curve

Bae, Whasoo,Kim, Ryeongah,Kim, Choongrak The Korean Statistical Society 2014 Communications for statistical applications and me Vol.21 No.1

In this paper we suggest an efficient method to estimate the distribution function using the Bezier curve, and compare it with existing methods by simulation studies. In addition, we suggest a robust version of cross-validation criterion to estimate the number of Bezier points, and showed that the proposed method is better than the existing methods based on simulation studies.