• Communications for Statistical Applications and Methods
• /
• v.20 no.4
• /
• pp.259-270
• /
• 2013

This paper considers a penalized composite quantile regression (CQR) that performs a variable selection in the linear model with grouped variables. An adaptive sup-norm penalized CQR (ASCQR) is proposed to select variables in a grouped manner; in addition, the consistency and oracle property of the resulting estimator are also derived under some regularity conditions. To improve the efficiency of estimation and variable selection, this paper suggests the two-stage penalized CQR (TSCQR), which uses the ASCQR to select relevant groups in the first stage and the adaptive lasso penalized CQR to select important variables in the second stage. Simulation studies are conducted to illustrate the finite sample performance of the proposed methods.

• The KIPS Transactions:PartD
• /
• v.8D no.4
• /
• pp.387-392
• /
• 2001

This Many software projects collect grouped failure data (failures in some failure interval or in variable time interval) rather than individual failure times or failure count data during the testing or operational phase. This paper presents the neural network (NN) modeling for grouped failure data that is able to predict cumulative failures in the variable future time. The two variant models of cascade-correlation learning (CasCor) algorithm are presented. Suggested models are compared with other well-known NN models and statistical software reliability growth models (SRGMs). Experimental results show that the suggested models show better predictability.

• The Transactions of the Korea Information Processing Society
• /
• v.7 no.12
• /
• pp.3821-3828
• /
• 2000

Many software projects collect grouped failure data (failures in some failure interval or in variable time interval) rather than individual failure times or failure count data during the testing or operational phase. This paper presents the neural network (NN) modeling that is dble to predict cumulative failures in the variable future time for grouped failure data. ANN's predictive ability can be affected by what it learns and in its ledming sequence. Eleven training regimes that represents the input-output of NN are considered. The best training regimes dre selected rJdsed on the next' step dvemge reldtive prediction error (AE) and normalized AE (NAE). The suggested NN models are compared with other well-known KN models and statistical software reliability growth models (SHGlvls) in order to evaluate performance, Experimental results show that the NN model with variable time interval information is necessary in order to predict cumulative failures in the variable future time interval.

• The Korean Journal of Applied Statistics
• /
• v.29 no.5
• /
• pp.961-975
• /
• 2016

The hierarchically penalized support vector machine (H-SVM) has been developed to perform simultaneous classification and input variable selection when input variables are naturally grouped or generated by factors. However, the H-SVM may suffer from estimation inefficiency because it applies the same amount of shrinkage to each variable without assessing its relative importance. In addition, when analyzing imbalanced data with uneven class sizes, the classification accuracy of the H-SVM may drop significantly in predicting minority class because its classifiers are undesirably biased toward the majority class. To remedy such problems, we propose the weighted adaptive H-SVM (WAH-SVM) method, which uses a adaptive tuning parameters to improve the performance of variable selection and the weights to differentiate the misclassification of data points between classes. Numerical results are presented to demonstrate the competitive performance of the proposed WAH-SVM over existing SVM methods.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2023.05a
• /
• pp.249-249
• /
• 2023

Spring drought forecasting in South Korea is essential due to the sknewness of rainfall which could lead to water shortage especially in spring when managed without prediction. Therefore, drought forecasting over South Korea was performed in the current study by thoroughly searching appropriate predictors from the lagged global climate variable, mean sea level pressure(MSLP), specifically in winter season for forecasting time lag. The target predictand defined as accumulated spring precipitation(ASP) was driven by the median of 93 weather stations in South Korea. Then, it was found that a number of points of the MSLP data were significantly cross-correlated with the ASP, and the points with high correlation were regionally grouped. The grouped variables with three regions: the Arctic Ocean (R1), South Pacific (R2), and South Africa (R3) were determined. The generalized linear model(GLM) was further applied for skewed marginal distribution in drought prediction. It was shown that the applied GLM presents reasonable performance in forecasting ASP. The results concluded that the presented regionalization of the climate variable, MSLP can be a good alternative in forecasting spring drought.

• The Journal of Korean Institute of Electromagnetic Engineering and Science
• /
• v.19 no.9
• /
• pp.1058-1064
• /
• 2008

In this paper, we proposed a new hybrid interference canceller using the adaptive filter with Grouped CMA(Constant Modulus Algorithm)-LMS(Least Mean Square) algorithm in the RF(Radio Frequency) repeater. The feedback signal generated from transmitter antenna to receiver antenna reduces the performance of the receiver system. The proposed interference canceller has better channel adaptive performance and a lower MSE(Mean Square Error) than conventional structure because it uses the cancellation method of Grouped CMA algorithm. This structure reduces the number of iterations fur the same MSE performance and hardware complexity compared to conventional nonlinear interference canceller. Namely, MSE values of the proposed algorithm were lower than those of LMS algorithm by 2.5 dB and 4 dB according to step sizes. And the proposed algorithm showed fast speed of convergence and similar MSE performance compared to VSS(Variable Step Size)-LMS algorithm.

• The Korean Journal of Applied Statistics
• /
• v.30 no.1
• /
• pp.135-145
• /
• 2017

Principal component analysis (PCA) describes the variation of multivariate data in terms of a set of uncorrelated variables. Since each principal component is a linear combination of all variables and the loadings are typically non-zero, it is difficult to interpret the derived principal components. Sparse principal component analysis (SPCA) is a specialized technique using the elastic net penalty function to produce sparse loadings in principal component analysis. When data are structured by groups of variables, it is desirable to select variables in a grouped manner. In this paper, we propose a new PCA method to improve variable selection performance when variables are grouped, which not only selects important groups but also removes unimportant variables within identified groups. To incorporate group information into model fitting, we consider a hierarchical lasso penalty instead of the elastic net penalty in SPCA. Real data analyses demonstrate the performance and usefulness of the proposed method.

• 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.

• ETRI Journal
• /
• v.31 no.5
• /
• pp.510-517
• /
• 2009

A high throughput parallel decoding method is developed for context-based adaptive variable length codes. In this paper, several new design ideas are devised and implemented for scalable parallel processing, a reduction in area, and a reduction in power requirements. First, simplified logical operations instead of memory lookups are used for parallel processing. Second, the codes are grouped based on their lengths for efficient logical operation. Third, up to M bits of the input stream can be analyzed simultaneously. For comparison, we designed a logical-operation-based parallel decoder for M=8 and a conventional parallel decoder. High-speed parallel decoding becomes possible with our method. In addition, for similar decoding rates (1.57 codes/cycle for M=8), our new approach uses 46% less chip area than the conventional method.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.5 no.11
• /
• pp.1959-1975
• /
• 2011

Adaptive Bit Loading (ABL) in Multiple-Input Multiple-Output Orthogonal Frequency-Division Multiplexing (MIMO-OFDM) is often used to achieve the desired Bit Error Rate (BER) performance in wireless systems. In this paper, we discuss some of the bit loading algorithms, compare them in terms of the BER performance, and present an effective and concise Adaptive Grouped Loading (AGL) algorithm. Furthermore, we propose a "weight factor" for loading algorithm to converge rapidly to the final solution for various data rate with variable Signal to Noise Ratio (SNR) gaps. In particular, we consider the bit loading in near optimal Singular Value Decomposition (SVD) based MIMO-OFDM system. While using SVD based system, the system requires perfect Channel State Information (CSI) of channel transfer function at the transmitter. This scenario of SVD based system is taken as an ideal case for the comparison of loading algorithms and to show the actual enhancement achievable by our AGL algorithm. Irrespective of the CSI requirement imposed by the mode of the system itself, ABL demands high level of feedback. Grouped Loading (GL) would reduce the feedback requirement depending upon the group size. However, this also leads to considerable degradation in BER performance. In our AGL algorithm, groups are formed with a number of consecutive sub-channels belonging to the same transmit antenna, with individual gains satisfying predefined criteria. Simulation results show that the proposed "weight factor" leads a loading algorithm to rapid convergence for various data rates with variable SNR gap values and AGL requires much lesser CSI compared to GL for the same BER performance.