• Journal of the Korean Data and Information Science Society
• /
• v.24 no.5
• /
• pp.1077-1088
• /
• 2013

High-dimensional data analysis arises from almost all scientific areas, evolving with development of computing skills, and has encouraged penalized estimations that play important roles in statistical learning. For the past years, various penalized estimations have been developed, and the least absolute shrinkage and selection operator (LASSO) proposed by Tibshirani (1996) has shown outstanding ability, earning the first place on the development of penalized estimation. In this paper, we first introduce a number of recent advances in high-dimensional data analysis using the LASSO. The topics include various statistical problems such as variable selection and grouped or structured variable selection under sparse high-dimensional linear regression models. Several unsupervised learning methods including inverse covariance matrix estimation are presented. In addition, we address further studies on new applications which may establish a guideline on how to use the LASSO for statistical challenges of high-dimensional data analysis.

• Asian-Australasian Journal of Animal Sciences
• /
• v.8 no.4
• /
• pp.365-368
• /
• 1995

Lactation curves of dairy cows were generated using three models, namely; incomplete gamma function (model 1), polynomial inverse function (model 2) and non-linear regression (model 3). Secondary milk yield data of 27 cows which had completed 6 lactations were used in this study. Milk yield records (once a week) throughout the lactation and from the first three months of lactation were fitted to the models. Estimation of total milk yield by model 3 using the data once a week throughout the lactation resulted in smaller % bias and standard error than those generated from model 1 and 2. But, model 2 was more accurate in predicting the 305-day milk yield equivalent closer to actual yields with smaller bias % and error using partial records up to 3 months. Also, model 2 was able to estimate the time to reach peak yield close to the actual data using partial records and model 2 could be used as a tool to advise farmers on appropriate feeding and management practices to be adopted.

• Journal of Korean Society for Quality Management
• /
• v.22 no.2
• /
• pp.89-97
• /
• 1994

We consider the linear calibration model: $y_1={\alpha}+{\beta}x_i+{\sigma}{\varepsilon}_i$, i = 1, ${\cdots}$, n, $y={\alpha}+{\beta}x+{\sigma}{\varepsilon}$ where ($y_1$, ${\cdots}$, $y_n$, y) stands for an observation vector, {$x_i$} fixed design vector, (${\alpha}$, ${\beta}$) vector of regression parameters, x unknown true value of interest and {${\varepsilon}_i$}, ${\varepsilon}$ are mutually uncorrelated measurement errors with zero mean and unit variance but otherwise unknown distributions. On the basis of simple small-sample low-noise approximation, we introduce a new method of comparing the mean squared errors of the various competing estimators of the true value x for finite sample size n. Then we show that a class of estimators including the classical and the inverse estimators are consistent and first-order efficient within the class of all regular consistent estimators irrespective of type of measurement errors.

• Journal of the Korea Furniture Society
• /
• v.27 no.3
• /
• pp.165-174
• /
• 2016

To modularize the joints of Hanok, the bending strengths of butterfly joints between pre-cut and manual member were compared. Structural size joints were manufactured and the length, width and thickness of each tenon were produced with different sizes. The ultimate load of pre-cut members was 2 times higher than that of manual members. Degree of anchorage for the joints on pre-cut member was also superior to that of manual member. By the F-test results, a great influence between ultimate load and sizes of tenon was found. In result of multiple regression analysis, the length and thickness of tenon were showed proportion relationships with the ultimate load, but the width of tenon was showed inverse proportion with the ultimate load. The results of this study can be used to identify the relationships among the major influence factors. Futhermore, it might be used as basic data for modularization the joints of Hanok.

• Journal of the Korean Statistical Society
• /
• v.28 no.2
• /
• pp.265-277
• /
• 1999

Classical and inverse estimation methods are two well known methods in statistical calibration problems. When there are outliers, both methods have large MSE's and could not estimate the input value correctly. We suggest median calibration estimation based on the LD-statistics. To investigate the robust performances, the influence function of the median calibration estimator is calculated and compared with other methods. When there are outliers in the response variables, the influence function is found to be bounded. In simulation studies, the MSE's for each calibration methods are compared. The estimated inputs as well as the performance of the influence functions are calculated.

• Journal of Korean Society of Water and Wastewater
• /
• v.7 no.2
• /
• pp.39-46
• /
• 1993

In this experimental study, it is concerned to develop a simple equation using jar-test results in order to predict the optimum dosage of coagulant, PAC(polyaluminum chloride). Considering the relationships with the reactions of coagulation and flocculation, the four independent variables (e.g. turbidity, temperature, pH and alkalinity) are selected out of many parameters and they are put into calculations to develop an equation by means of multi-regression method. As the result, the dosing rate of PAC is proportional to turbidity, pH and alkalinity, but in inverse to temperature. And the developed equation is as follow, $$D_c=\frac{3.2{\cdot}T^{0.37}{\cdot}A^{0.04}{\cdot}P^{0.5}}{t^{0.1}},\;(R^2=0.9443)$$ And also, comparing between the estimated value from the equation and the real dosing rate in the plant, Kwangam and Tdukdo, during 1988~1991, it is represented an agreement having a relative error of 16.4%, 17.8%, respectively.

• Structural Engineering and Mechanics
• /
• v.70 no.1
• /
• pp.1-11
• /
• 2019

This study deals with a method based on the modified bivariate gamma function for reconstructions of dynamic behavior of delaminated composite plates subjected to impact loads. The proposed bivariate gamma function is associated with micro-genetic algorithms, which is capable of solving inverse problems to determine the stiffness reduction associated with delamination. From computing the unknown parameters, it is possible for the entire dynamic response data to develop a prediction model of the dynamic response through a regression analysis based on the measurement data. The validity of the proposed method was verified by comparing with results employing a higher-order finite element model. Parametric results revealed that the proposed method can reconstruct dynamic responses and the stiffness reduction of delaminated composite plates can be investigated for different measurements and loading locations.

• Journal of KIISE:Software and Applications
• /
• v.34 no.1
• /
• pp.70-78
• /
• 2007

The performance of face recognition is greatly affected by the illumination effect because intra-person variation under different lighting conditions can be much bigger than the inter-person variation. In this paper, we propose an illumination robust face recognition by separating identity factor and illumination factor using the symmetric bilinear models. The translation procedure in the bilinear model requires a repetitive computation of matrix inverse operation to reach the identity and illumination factors. Sometimes, this computation may result in a nonconvergent case when the observation has an noisy information. To alleviate this situation, we suggest a ridge regressive bilinear model that combines the ridge regression into the bilinear model. This combination provides some advantages: it makes the bilinear model more stable by shrinking the range of identity and illumination factors appropriately, and it improves the recognition performance by reducing the insignificant factors effectively. Experiment results show that the ridge regressive bilinear model outperforms significantly other existing methods such as the eigenface, quotient image, and the bilinear model in terms of the recognition rate under a variety of illuminations.

• Journal of Biomedical Engineering Research
• /
• v.25 no.6
• /
• pp.605-609
• /
• 2004

Blood pressure (BP) is one of the important physiological parameters for diagnosing cardiovascula diseases by means of noninvasive method. Existing noninvasive methods for measuring arterial BP have to use cuff and difficult in measuring arterial BP continuously. Systolic blood pressure (SBP) and pulse transit time (PTT) have a kind of inverse relationship. We acquired PTT data when subjects were in relaxation and also after exercise. We performed the linear regression analysis for making the regression equations for each subject and the regression equation for all subjects. We compared the estimated SBP with the measured SBP to check the accuracy of our regression equations. From the result, the regression equations for each subject was appropriate according to the American National Standards Institute of the Association of the Advancement of Medical Instrument (ANSI/AAMI) which says that BP devices should have ±5mmHg mean of error and 8mmHg standard deviation of error. However, the regression equation for all subjects was not proper to ANSI/AAMI recommendation. The result means that, without cuff, we can continuously estimate each subject's SBP through PTT and indivisual calibration.

• Atmosphere
• /
• v.24 no.1
• /
• pp.101-110
• /
• 2014

This study estimates and evaluates the daily January temperature from 2003 to 2012 with 30 m-resolution over South Korea, using a modified Parameter-elevation Regression on Independent Slopes Model (K-PRISM). Several factors in K-PRISM are also adjusted to 30 m grid spacing and daily time scales. The performance of K-PRISM is validated in terms of bias, root mean square error (RMSE), and correlation coefficient (Corr), and is then compared with that of inverse distance weighting (IDW) and hypsometric methods (HYPS). In estimating the temperature over Jeju island, K-PRISM has the lowest bias (-0.85) and RMSE (1.22), and the highest Corr (0.79) among the three methods. It captures the daily variation of observation, but tends to underestimate due to a high-discrepancy in mean altitudes between the observation stations and grid points of the 30 m topography. The temperature over South Korea derived from K-PRISM represents a detailed spatial pattern of the observed temperature, but generally tends to underestimate with a mean bias of -0.45. In bias terms, the estimation ability of K-PRISM differs between grid points, implying that care should be taken when dealing with poor skill area. The study results demonstrate that K-PRISM can reasonably estimate 30 m-resolution temperature over South Korea, and reflect topographically diverse signals with detailed structure features.