• Quantitative Bio-Science
• /
• v.37 no.2
• /
• pp.133-141
• /
• 2018

Graphical lasso is one of the most popular methods to estimate a sparse precision matrix, which is an inverse of a covariance matrix. The objective function of graphical lasso imposes an ${\ell}_1$-penalty on the (vectorized) precision matrix, where a tuning parameter controls the strength of the penalization. The selection of the tuning parameter is practically and theoretically important since the performance of the estimation depends on an appropriate choice of tuning parameter. While information criteria (e.g. AIC, BIC, or extended BIC) have been widely used, they require an asymptotically unbiased estimator to select optimal tuning parameter. Thus, the biasedness of the ${\ell}_1$-regularized estimate in the graphical lasso may lead to a suboptimal tuning. In this paper, we propose a two-staged bias-correction procedure for the graphical lasso, where the first stage runs the usual graphical lasso and the second stage reruns the procedure with an additional constraint that zero estimates at the first stage remain zero. Our simulation and real data example show that the proposed bias correction improved on both edge recovery and estimation error compared to the single-staged graphical lasso.

• Journal of Korean Society for Quality Management
• /
• v.19 no.2
• /
• pp.17-39
• /
• 1991

In many application of response surface methodology, good estimation of the derivatives of the response function may be as important or perhaps more important than estimation of mean response. Using a graphical method, we have studied the effect of outliers, missing observations, and design augmentation with respect to the slope estimation in the response surf ace designs.

• Management Science and Financial Engineering
• /
• v.2 no.1
• /
• pp.103-122
• /
• 1996

This paper presents an easily used graphical procedure for simultaneous estimation of the index, skewness, scale, and location parameters of the stable laws. First, the index $\alpha$ and skewness $\beta$ are estimated through the joint use of a tail length statistic $\widetilde{K_t}$ and a skewness statistic $\widetilde{K_s}$, both of which are functions of order statistics. Next, the function of order statistics needed for estimation of scale $\sigma$ and location $\mu$ are determined from a nomogram indexed on the estimates of $\alpha$ and $\beta$. Some applications and examples are provided.

• Communications for Statistical Applications and Methods
• /
• v.4 no.2
• /
• pp.359-365
• /
• 1997

Unusual observations can greatly influence the results of least wquares estimation. I propose graphical methods which can detect the influential observations.

• Journal of the Korean Society for Precision Engineering
• /
• v.13 no.5
• /
• pp.104-114
• /
• 1996

For the improvement of productivity, the reduction of cost and time for manufacturing is mandatory, especially in the field of electromic industry. The study is concerned with a practical means of systematic assistance to formability estimation and selection of reliable design specification for electronic sheet metal parts. The objective of this research work is to develop a simulation system which hops to analyze the target processes with the finite element method and to acquire available design data quickly and exactly. The simulation system developed in the study consists of design verification, selection of optimal combination of parameters, knowledge acquisition and graphical user interface(GUI). Design verification is automatically carried out by using the finite element method. A data base management system and nomograms are utilized for knowledge acquisition. The developed system has been applied to some major sheet metal forming operations such as flanging, embossing, bending and blanking. According to the simulated results, the validation of the target processes has been confirmend. Analysis data, estimation rules of formability and graphical representation of the analysis have been employed for the designer's understanding and evaluation, thus providing a practical means of robust design and evaluation of forma- bility for producing electronic sheet metal parts.

• Journal of Korean Society for Quality Management
• /
• v.23 no.2
• /
• pp.68-79
• /
• 1995

Good estimation of the slopes of the mixture response function may be important as well as estimation of mean mixture response. It is possible to evaluate and compare several mixture designs with respect to the slope. A graphical method is proposed that allows us to evaluate a given design's support for the fitted model in terms of slope variance. We can plot variances of slopes along Cox direction according to existence of restriction of simplex region when comparing several different mixture designs.

• Journal of the Korean Data and Information Science Society
• /
• v.25 no.4
• /
• pp.881-891
• /
• 2014

Graphical models can be used as an intuitive tool for modeling a complex stochastic system with a large number of variables related each other because the conditional independence between random variables can be visualized as a network. Graphical least absolute shrinkage and selection operator (LASSO) is considered to be effective in avoiding overfitting in the estimation of Gaussian graphical models for high dimensional data. In this paper, we consider the model selection problem in graphical LASSO. Particularly, we compare various model selection criteria via simulations and analyze a real financial data set.

• Journal of the Korean Society for Precision Engineering
• /
• v.19 no.10
• /
• pp.107-113
• /
• 2002

To monitor the torque of an induction motor using current, rotating speed has been measured and used to calculate the slip angular velocity. Additional sensor, however, can cause extra expense and trouble. In this paper, a new algorithm is proposed to monitor the torque of vector controlled induction motor without any speed measuring sensor. Only stator current is measured to estimate the magnetizing current which is used to calculate flux linkage, rotor velocity and motor torque. Graphical programming is used to implement the suggested algorithm and to monitor the torque of an induction motor in real time. To solve the fluctuation problem of estimated torque caused from instantly varying rotating speed of an induction motor, the rotating speed is reconstructed based on the measured current signals. From several experiments, the proposed method shows a good estimation of the motor torque under the normal rotational speed.

• Proceedings of the Korean Statistical Society Conference
• /
• 2003.05a
• /
• pp.213-217
• /
• 2003

In this paper we discuss graphical and diagnostic methods for logistic regression, in which the response is the number of successes in a fixed number of trials.

• Communications for Statistical Applications and Methods
• /
• v.11 no.2
• /
• pp.275-285
• /
• 2004

We explore classification analysis using graphical methods such as sliced inverse regression and sliced average variance estimation based on dimension reduction. Some useful information about classification analysis are obtained by sliced inverse regression and sliced average variance estimation through dimension reduction. Two examples are illustrated, and classification rates by sliced inverse regression and sliced average variance estimation are compared with those by discriminant analysis and logistic regression.