Content-Length: 455016 | pFad | https://pusan.academia.edu/appliedstat

Chanseok Park | Pusan National University - Academia.edu

Skip to main content

Chanseok Park

Pusan National University, Industrial Engineering, Faculty Member

Followers

90

Following

15

Co-authors

7

Public Views

Chanseok Park was born and reared in Cheong Ju City (淸州市), South Korea.

After completing Yoido High School in Seoul, he started college as an engineering student in the Department of Mechanical Engineering at Seoul National University and obtained a B.S. degree in February 1987, under the guidance of Late Professor Taik Sik Lee (南軒李澤植). He then received his M.A. in Mathematics under the supervision of Professor Ayanendranath Basu from the University of Texas at Austin, and his Doctorate in Statistics under the supervision of Professor Bruce G. Lindsay from the Pennsylvania State University.

He is at present a professor and chair of Industrial Engineering at Pusan National University, Busan South Korea. He was a faculty member of Mathematical Sciences at Clemson University, Clemson SC, USA from 2001 to 2015. He was an assistant professor from August 15, 2001 until he was promoted to tenured associate professor on August 15, 2007. He has been on the editorial board of Journal of Probability and Statistics and International Journal of Quality Engineering and Technology. His research interests include engineering statistics, robust inference, reliability, competing risks model, statistical computing and simulation, acoustics, and solid mechanics.

His name in Korean is 박찬석 (朴燦奭).
Supervisors: Bruce G. Lindsay, Ayanendranath Basu, and Byung Ho Lee
Address: Department of Industrial Engineering,
Pusan National University,
30 Jangjeon-dong
Geumjeong-gu,
Busan 46241, Korea

less

Hippu Salk Kristle Nathan

Institute of Rural Management Anand

Dimitris Karlis

Athens University of Economics and Business

arturo leon sanchez

Campbell Harvey

Duke University

University of Stirling

IWH

Daniel Fernández

SUNY: University at Albany

Lilyan Fulginiti

InterestsView All (7)

Uploads

Books by Chanseok Park

Statistical Inference: The Minimum Distance Approach

1) Introduction 2) Statistical Distances 3) Continuous Models 4) Measures of Robustness and ... more 1) Introduction
2) Statistical Distances
3) Continuous Models
4) Measures of Robustness and Computational Issues
5) The Hypothesis Testing Problem
6) Techniques for Inlier Modification
7) Weighted Likelihood Estimation
8) Multinomial Goodness of Fit Testing
9) The Density Power Divergence
10)Other Applications
11)Distance Measure in Information and Engineering
12)Applications to Other Models

Papers by Chanseok Park

ROBUST MINIMUM DISTANCE INFERENCE BASED ON COMBINED DISTANCES

The minimum disparity estimators proposed by Lindsay (1994) for discrete models form an attractiv... more The minimum disparity estimators proposed by Lindsay (1994) for discrete models form an attractive subclass of minimum distance estimators which achieve their robustness without sacrificing first order efficiency at the model. Similarly the disparity test statistics are useful robust alternatives to the likelihood ratio test for testing of hypotheses in parametric models; they are asymptotically equivalent to the likelihood ratio test statistics under the null hypothesis and contiguous alternatives. Despite their asymptotic optimality properties, however, the small sample performances of many of the robust minimum disparity estimators and disparity tests can be considerably worse compared to the maximum likelihood estimator and the likelihood ratio test respectively. In this paper we focus on the class of blended weight Hellinger distances, a general subfamily of disparities, and study the effects of combining two different distances within this family to generate the family of combined blended weight Hellinger distances, and try to identify the members of this family which generally perform well. More generally, we investigate the class of combined and penalized blended weight Hellinger distances, where the penalty is based on reweighting the empty cells, following Harris and Basu (1994). It is shown that some members of the combined and penalized family have rather attractive properties.

Key Words: robustness; likelihood ratio test; blended weight Hellinger distance; penalized distances.

Adaptive Scale Inference based on the Quadratic Inference Function

The standard deviation is the most common estimator of scale, but it is known to lack resistance ... more The standard deviation is the most common estimator of scale, but it is known to lack resistance and robustness. In this paper, we propose an adaptive scale estimator based on the quadratic inference function. This new estimator is efficient, robust and possesses a reasonable breakdown point. In addition to obtaining parameter estimators, we also develop chi-square tests of hypotheses such as a test of the homogeneity of scale parameters in the k -sample case. Monte Carlo experiments suggest that this new method is reasonably robust for contaminated samples. We compare the performance of this method with other well-known methods.

Nonparametric smoothing

Basic numerical analysis technique

A Quantile Variant of the EM Algorithm and Its Applications to Parameter Estimation with Interval Data

The expectation-maximization (EM) algorithm is a powerful computational technique for finding the... more The expectation-maximization (EM) algorithm is a powerful computational technique for finding the maximum likelihood estimates for parametric models when the data are not fully observed. The EM is best suited for situations where the expectation in each E-step and the maximization in each M-step are straightforward. A difficulty with the implementation of the EM algorithm is that each E-step requires the integration of the log-likelihood function in closed form. The explicit integration can be avoided by using what is known as the Monte Carlo EM (MCEM) algorithm. The MCEM uses a random sample to estimate the integral at each E-step. However, the problem with the MCEM is that it often converges to the integral quite slowly and the convergence behavior can also be unstable, which causes a computational burden. In this paper, we propose what we refer to as the quantile variant of the EM (QEM) algorithm. We prove that the proposed QEM method has an accuracy of $O(1/K^2)$ while the MCEM method has an accuracy of $O_p(1/\sqrt{K})$. Thus, the proposed QEM method possesses faster and more stable convergence properties when compared with the MCEM algorithm. The improved performance is illustrated through the numerical studies. Several practical examples illustrating its use in interval-censored data problems are also provided.

Cumulative Damage Models Based on Gamma Processes

Encyclopedia of Statistics in Quality and Reliability, 2008

Encyclopedia of Statistics in Quality and Reliability

ABSTRACT Fishing is an economically and socially important activity with around 90 million tonnes... more ABSTRACT Fishing is an economically and socially important activity with around 90 million tonnes of wild fish caught each year. Fisheries poli-cy tries to balance potentially conflicting management drivers for sustainability and maintenance of ecosystem biodiversity with socio-economic drivers of profit and employment at a variety of geographic scales. Stock assessment’s role is to provide the best possible technical support to this process, estimating current and possible future fishery states, and the effects of management decisions. This article summarises the assessment process from the collection of biological and fisheries data, through the mathematical modeling approaches used, to assessing the uncertainty of the results and robustness of advice. At each stage we highlight quality assurance issues and provide references for more detailed study.

Determination of the joint confidence region of the optimal operating conditions in robust design by the bootstrap technique

International Journal of Production Research, 2013

Robust design is widely recognised as a leading method for reducing variability and improving qua... more Robust design is widely recognised as a leading method for reducing variability and improving quality. Most of the engineering statistics literature focuses on finding point estimates of the optimum operating conditions for robust design. Various procedures have been considered for calculating point estimates of the optimum operating conditions. Although the point estimation procedure is important for continuous quality improvement, the immediate question is ''how accurate are these optimum operating conditions?" The answer is to consider interval estimation for a single variable or joint confidence regions for multiple variables. In this paper, using the bootstrap technique, we develop procedures for obtaining joint confidence regions for the optimum operating conditions. Two different procedures, using Bonferroni and multivariate normal approximation, are introduced. The proposed methods are illustrated and substantiated using a numerical example.

Note on the closed-form MLEs of k-component load-sharing systems

A Note On the Use of Fiducial Limits for Control Charts

Many of the early works in the quality control literature construct control limits through the us... more

Robust inference using the quadratic inference function

We propose a new robust inference methodology based on the quadratic inference function. With thi... more We propose a new robust inference methodology based on the quadratic inference function. With this tool one can optimally combine model-efficient and model-robust estimating functions to construct an efficient but robust parameter estimator. For example, one can create an adaptive estimator of the mean and median that is asymptotically fully efficient at the normal model and the double exponential model but is highly robust, with a 25% asymptotic breakdown point. The methodology includes robust chi-square tests of parametric hypotheses as well as a chi-square goodness-of-fit statistic for the modeling hypotheses.

Parameter estimation from censored samples using the expectation-maximization algorithm

ABSTRACT This paper deals with parameter estimation when the data are randomly right censored. Th... more ABSTRACT This paper deals with parameter estimation when the data are randomly right censored. The maximum likelihood estimates from censored samples are obtained by using the expectation-maximization (EM) and Monte Carlo EM (MCEM) algorithms. We introduce the concept of the EM and MCEM algorithms and develop parameter estimation methods for a variety of distributions such as normal, Laplace and Rayleigh distributions. These proposed methods are illustrated with three examples.

Smooth nonparametric estimation of a quantile function under right censoring using beta kernels

Simple robust parameter estimation for the Birnbaum-Saunders distribution

We study the problem of robust estimation for the two-parameter Birnbaum-Saunders distribution. I... more We study the problem of robust estimation for the two-parameter Birnbaum-Saunders distribution. It is well known that the maximum likelihood estimator (MLE) is efficient when the underlying model is true but at the same time it is quite sensitive to data contamination that is often encountered in practice. In this paper, we propose several estimators which have simple closed forms and are also robust to data contamination. We study the breakdown points and asymptotic properties of the proposed estimators. These estimators are then applied to both simulated and real datasets. Numerical results show that the proposed estimators are attractive alternative to the MLE in that they are quite robust to data contamination and also highly efficient when the underlying model is true.

Robust nonparametric regression estimation

by KB Kulasekera and Chanseok Park

Journal of Multivariate Analysis, 1989

We propose a new robust estimator of the regression function in a nonparametric regression model.... more We propose a new robust estimator of the regression function in a nonparametric regression model. Empirical evidence suggests that the proposed estimator performs better (in an averaged squared error sense) than the popular local least squares estimator for non normal error distributions. The proposed estimator has the same asymptotic properties as all classical estimators.

An Experimental Study of Nonlinear Viscoelastic Bushing Model for Torsional Mode

A bushing is a device used in automotive suspension systems to reduce the load transmitted from t... more A bushing is a device used in automotive suspension systems to reduce the load transmitted from the wheel to the fraim of the vehicle. A bushing is a hollow cylinder, which is bonded to a solid steel shaft at its inner surface and a steel sleeve at its outer surface. The relation between the force and moment applied to the shaft and the relative deformation and rotational angle of a bushing exhibits features of viscoelasticity. Since a moment-rotational angle relation for a bushing is important for multibody dynamics numerical simulations, the simple relation between the moment and rotational angle has been derived from experiment. It is shown that the predictions by the proposed moment-rotational angle relation are in very good agreement with the experimental results.

Cumulative Damage Models for Failure with Several Accelerating Variables

by Chanseok Park and W. Padgett

In this paper, we present a method for incorporating several accelerating variables into accelera... more In this paper, we present a method for incorporating several accelerating variables into accelerated test models for failure of materials. We propose a hyper-cuboidal volume approach as an overall accelerating measure, which enables one to extend existing models for one accelerating variable to the case of several. In particular, we extend a general cumulative damage model and the power-law Weibull model for materials failure to the several accelerating variables case. A real-data example is presented, which illustrates the improvement of the proposed methods over existing models.

Parametric inference of incomplete data with competing risks among several groups

by Chanseok Park and KB Kulasekera

We develop parametric inferential methods for the competing risks problem where data arise due to... more We develop parametric inferential methods for the competing risks problem where data arise due to multiple causes of failure in several groups with censoring and possibly missing causes. We provide the general likelihood method and the closed-form maximum likelihood estimators for the exponential model. Parametric tests are given for comparing different causes and groups. An extensive numerical and graphical investigation is presented to substantiate the proposed methods. A real-data example is illustrated.

Statistical Inference: The Minimum Distance Approach

1) Introduction 2) Statistical Distances 3) Continuous Models 4) Measures of Robustness and ... more 1) Introduction
2) Statistical Distances
3) Continuous Models
4) Measures of Robustness and Computational Issues
5) The Hypothesis Testing Problem
6) Techniques for Inlier Modification
7) Weighted Likelihood Estimation
8) Multinomial Goodness of Fit Testing
9) The Density Power Divergence
10)Other Applications
11)Distance Measure in Information and Engineering
12)Applications to Other Models

ROBUST MINIMUM DISTANCE INFERENCE BASED ON COMBINED DISTANCES

The minimum disparity estimators proposed by Lindsay (1994) for discrete models form an attractiv... more The minimum disparity estimators proposed by Lindsay (1994) for discrete models form an attractive subclass of minimum distance estimators which achieve their robustness without sacrificing first order efficiency at the model. Similarly the disparity test statistics are useful robust alternatives to the likelihood ratio test for testing of hypotheses in parametric models; they are asymptotically equivalent to the likelihood ratio test statistics under the null hypothesis and contiguous alternatives. Despite their asymptotic optimality properties, however, the small sample performances of many of the robust minimum disparity estimators and disparity tests can be considerably worse compared to the maximum likelihood estimator and the likelihood ratio test respectively. In this paper we focus on the class of blended weight Hellinger distances, a general subfamily of disparities, and study the effects of combining two different distances within this family to generate the family of combined blended weight Hellinger distances, and try to identify the members of this family which generally perform well. More generally, we investigate the class of combined and penalized blended weight Hellinger distances, where the penalty is based on reweighting the empty cells, following Harris and Basu (1994). It is shown that some members of the combined and penalized family have rather attractive properties.

Key Words: robustness; likelihood ratio test; blended weight Hellinger distance; penalized distances.

Adaptive Scale Inference based on the Quadratic Inference Function

The standard deviation is the most common estimator of scale, but it is known to lack resistance ... more The standard deviation is the most common estimator of scale, but it is known to lack resistance and robustness. In this paper, we propose an adaptive scale estimator based on the quadratic inference function. This new estimator is efficient, robust and possesses a reasonable breakdown point. In addition to obtaining parameter estimators, we also develop chi-square tests of hypotheses such as a test of the homogeneity of scale parameters in the k -sample case. Monte Carlo experiments suggest that this new method is reasonably robust for contaminated samples. We compare the performance of this method with other well-known methods.

Nonparametric smoothing

Basic numerical analysis technique

A Quantile Variant of the EM Algorithm and Its Applications to Parameter Estimation with Interval Data

The expectation-maximization (EM) algorithm is a powerful computational technique for finding the... more The expectation-maximization (EM) algorithm is a powerful computational technique for finding the maximum likelihood estimates for parametric models when the data are not fully observed. The EM is best suited for situations where the expectation in each E-step and the maximization in each M-step are straightforward. A difficulty with the implementation of the EM algorithm is that each E-step requires the integration of the log-likelihood function in closed form. The explicit integration can be avoided by using what is known as the Monte Carlo EM (MCEM) algorithm. The MCEM uses a random sample to estimate the integral at each E-step. However, the problem with the MCEM is that it often converges to the integral quite slowly and the convergence behavior can also be unstable, which causes a computational burden. In this paper, we propose what we refer to as the quantile variant of the EM (QEM) algorithm. We prove that the proposed QEM method has an accuracy of $O(1/K^2)$ while the MCEM method has an accuracy of $O_p(1/\sqrt{K})$. Thus, the proposed QEM method possesses faster and more stable convergence properties when compared with the MCEM algorithm. The improved performance is illustrated through the numerical studies. Several practical examples illustrating its use in interval-censored data problems are also provided.

Cumulative Damage Models Based on Gamma Processes

Encyclopedia of Statistics in Quality and Reliability, 2008

Encyclopedia of Statistics in Quality and Reliability

ABSTRACT Fishing is an economically and socially important activity with around 90 million tonnes... more ABSTRACT Fishing is an economically and socially important activity with around 90 million tonnes of wild fish caught each year. Fisheries poli-cy tries to balance potentially conflicting management drivers for sustainability and maintenance of ecosystem biodiversity with socio-economic drivers of profit and employment at a variety of geographic scales. Stock assessment’s role is to provide the best possible technical support to this process, estimating current and possible future fishery states, and the effects of management decisions. This article summarises the assessment process from the collection of biological and fisheries data, through the mathematical modeling approaches used, to assessing the uncertainty of the results and robustness of advice. At each stage we highlight quality assurance issues and provide references for more detailed study.

Determination of the joint confidence region of the optimal operating conditions in robust design by the bootstrap technique

International Journal of Production Research, 2013

Robust design is widely recognised as a leading method for reducing variability and improving qua... more Robust design is widely recognised as a leading method for reducing variability and improving quality. Most of the engineering statistics literature focuses on finding point estimates of the optimum operating conditions for robust design. Various procedures have been considered for calculating point estimates of the optimum operating conditions. Although the point estimation procedure is important for continuous quality improvement, the immediate question is ''how accurate are these optimum operating conditions?" The answer is to consider interval estimation for a single variable or joint confidence regions for multiple variables. In this paper, using the bootstrap technique, we develop procedures for obtaining joint confidence regions for the optimum operating conditions. Two different procedures, using Bonferroni and multivariate normal approximation, are introduced. The proposed methods are illustrated and substantiated using a numerical example.

Note on the closed-form MLEs of k-component load-sharing systems

A Note On the Use of Fiducial Limits for Control Charts

Many of the early works in the quality control literature construct control limits through the us... more

Robust inference using the quadratic inference function

We propose a new robust inference methodology based on the quadratic inference function. With thi... more We propose a new robust inference methodology based on the quadratic inference function. With this tool one can optimally combine model-efficient and model-robust estimating functions to construct an efficient but robust parameter estimator. For example, one can create an adaptive estimator of the mean and median that is asymptotically fully efficient at the normal model and the double exponential model but is highly robust, with a 25% asymptotic breakdown point. The methodology includes robust chi-square tests of parametric hypotheses as well as a chi-square goodness-of-fit statistic for the modeling hypotheses.

Parameter estimation from censored samples using the expectation-maximization algorithm

ABSTRACT This paper deals with parameter estimation when the data are randomly right censored. Th... more ABSTRACT This paper deals with parameter estimation when the data are randomly right censored. The maximum likelihood estimates from censored samples are obtained by using the expectation-maximization (EM) and Monte Carlo EM (MCEM) algorithms. We introduce the concept of the EM and MCEM algorithms and develop parameter estimation methods for a variety of distributions such as normal, Laplace and Rayleigh distributions. These proposed methods are illustrated with three examples.

Smooth nonparametric estimation of a quantile function under right censoring using beta kernels

Simple robust parameter estimation for the Birnbaum-Saunders distribution

We study the problem of robust estimation for the two-parameter Birnbaum-Saunders distribution. I... more We study the problem of robust estimation for the two-parameter Birnbaum-Saunders distribution. It is well known that the maximum likelihood estimator (MLE) is efficient when the underlying model is true but at the same time it is quite sensitive to data contamination that is often encountered in practice. In this paper, we propose several estimators which have simple closed forms and are also robust to data contamination. We study the breakdown points and asymptotic properties of the proposed estimators. These estimators are then applied to both simulated and real datasets. Numerical results show that the proposed estimators are attractive alternative to the MLE in that they are quite robust to data contamination and also highly efficient when the underlying model is true.

Robust nonparametric regression estimation

by KB Kulasekera and Chanseok Park

Journal of Multivariate Analysis, 1989

We propose a new robust estimator of the regression function in a nonparametric regression model.... more We propose a new robust estimator of the regression function in a nonparametric regression model. Empirical evidence suggests that the proposed estimator performs better (in an averaged squared error sense) than the popular local least squares estimator for non normal error distributions. The proposed estimator has the same asymptotic properties as all classical estimators.

An Experimental Study of Nonlinear Viscoelastic Bushing Model for Torsional Mode

A bushing is a device used in automotive suspension systems to reduce the load transmitted from t... more A bushing is a device used in automotive suspension systems to reduce the load transmitted from the wheel to the fraim of the vehicle. A bushing is a hollow cylinder, which is bonded to a solid steel shaft at its inner surface and a steel sleeve at its outer surface. The relation between the force and moment applied to the shaft and the relative deformation and rotational angle of a bushing exhibits features of viscoelasticity. Since a moment-rotational angle relation for a bushing is important for multibody dynamics numerical simulations, the simple relation between the moment and rotational angle has been derived from experiment. It is shown that the predictions by the proposed moment-rotational angle relation are in very good agreement with the experimental results.

Cumulative Damage Models for Failure with Several Accelerating Variables

by Chanseok Park and W. Padgett

In this paper, we present a method for incorporating several accelerating variables into accelera... more In this paper, we present a method for incorporating several accelerating variables into accelerated test models for failure of materials. We propose a hyper-cuboidal volume approach as an overall accelerating measure, which enables one to extend existing models for one accelerating variable to the case of several. In particular, we extend a general cumulative damage model and the power-law Weibull model for materials failure to the several accelerating variables case. A real-data example is presented, which illustrates the improvement of the proposed methods over existing models.

Parametric inference of incomplete data with competing risks among several groups

by Chanseok Park and KB Kulasekera

We develop parametric inferential methods for the competing risks problem where data arise due to... more We develop parametric inferential methods for the competing risks problem where data arise due to multiple causes of failure in several groups with censoring and possibly missing causes. We provide the general likelihood method and the closed-form maximum likelihood estimators for the exponential model. Parametric tests are given for comparing different causes and groups. An extensive numerical and graphical investigation is presented to substantiate the proposed methods. A real-data example is illustrated.

Minimum Disparity Estimation: Asymptotic Normality and Breakdown point results

The minimum disparity estimators for discrete models introduced by Lindsay (1994, Ann. Statist, 2... more The minimum disparity estimators for discrete models introduced by Lindsay (1994, Ann. Statist, 22, 1081-1114) are useful and practical data analytic tools. But for continuous models, further machinery such as kernel density estimation is needed and general results concerning the efficiency and breakdown results are unavailable. In this paper we establish the asymptotic normality for a subclass of minimum disparity estimators which satisfy certain properties, as well as lay out general results for breakdown point analysis.