• 한국지능시스템학회:학술대회논문집
• /
• 한국퍼지및지능시스템학회 2004년도 추계학술대회 학술발표 논문집 제14권 제2호
• /
• pp.313-316
• /
• 2004

Model selection is the process that sets up the regularization parameter in the support vector machine or regularization network by using the external methods such as general cross validation or L-curve criterion. This paper suggests that the regularization parameter can be obtained simultaneously within the learning process of neural networks without resort to separate selection methods. In this paper, extended kernel method is introduced. The relationship between regularization parameter and the bias term in the extended kernel is established. Experimental results show the effectiveness of the new model selection method.

• Communications for Statistical Applications and Methods
• /
• 제24권6호
• /
• pp.651-662
• /
• 2017

We propose an alternative procedure to select penalty parameter in $L_1$ penalized robust regression. This procedure is based on marginalization of prior distribution over the penalty parameter. Thus, resulting objective function does not include the penalty parameter due to marginalizing it out. In addition, its estimating algorithm automatically chooses a penalty parameter using the previous estimate of regression coefficients. The proposed approach bypasses cross validation as well as saves computing time. Variable-wise penalization also performs best in prediction and variable selection perspectives. Numerical studies using simulation data demonstrate the performance of our proposals. The proposed methods are applied to Boston housing data. Through simulation study and real data application we demonstrate that our proposals are competitive to or much better than cross-validation in prediction, variable selection, and computing time perspectives.

• 한국소음진동공학회:학술대회논문집
• /
• 한국소음진동공학회 2004년도 춘계학술대회논문집
• /
• pp.882-886
• /
• 2004

In this section, the proposed parameter selection procedure is applied to two example problems, one is the plate example given in section 2.2 and the other is a cover structure of hard disk drive (HDD).

• Journal of the Korean Statistical Society
• /
• 제31권1호
• /
• pp.93-107
• /
• 2002

Random effects generalised linear models are useful for analysing clustered count data in which responses are usually correlated. We propose a Bayesian approach to parameter estimation and variable selection in random effects generalised linear models for count data. A simple Gibbs sampling algorithm for parameter estimation is presented and a simple and efficient variable selection is done by using the Gibbs outputs. An illustrative example is provided.

• Journal of the Korean Statistical Society
• /
• 제32권2호
• /
• pp.151-162
• /
• 2003

In this paper, we propose two selection procedures for selecting populations better than a control population. The bestness is defined in terms of location parameter. One of the procedures is based on two-sample linear rank statistics whereas the other one is based on a comparatively simple statistic, and is useful when testing time is expensive so that an early termination of an experiment is desirable. The proposed selection procedures are seen to be strongly monotone. Performance of the proposed procedures is assessed through simulation study.

• Journal of the Korean Statistical Society
• /
• 제27권4호
• /
• pp.397-405
• /
• 1998

We develop parametric bootstrap model selection criteria in an example to fit a random sample to either a general normal distribution or a normal distribution with prespecified mean. We apply the bootstrap methods in two ways; one considers the direct substitution of estimated parameter for the unknown parameter, and the other focuses on the bias correction. These bootstrap model selection criteria are compared with AIC. We illustrate that all the selection rules reduce to the one sample t-test, where the cutoff points converge to some certain points as the sample size increases.

• Journal of Astronomy and Space Sciences
• /
• 제36권2호
• /
• pp.97-103
• /
• 2019

An automated signal-acquisition method for the NASA's space geodesy satellite laser ranging (SGSLR) system is described as a selection of two system parameters with specified probabilities. These parameters are the correlation parameter: the minimum received pulse number for a signal-acquisition and the frame time: the minimum time for the correlation parameter. The probabilities specified are the signal-detection and false-acquisition probabilities to distinguish signals from background noise. The steps of parameter selection are finding the minimum set of values by fitting a curve and performing a graph-domain approximation. However, this selection method is inefficient, not only because of repetition of the entire process if any performance values change, such as the signal and noise count rate, but also because this method is dependent upon system specifications and environmental conditions. Moreover, computation is complicated and graph-domain approximation can introduce inaccuracy. In this study, a new method is proposed to select the parameters via a conditional equation derived from characteristics of the signal-detection and false-acquisition probabilities. The results show that this method yields better efficiency and robustness against changing performance values with simplicity and accuracy and can be easily applied to other satellite laser ranging (SLR) systems.

• Journal of the Korean Data and Information Science Society
• /
• 제26권3호
• /
• pp.549-559
• /
• 2015

MDPDE는 이상치에 강건한 성질을 가진 추정량으로써 최대우도추정량의 대안으로 많은 연구자들에 의해 연구되어 왔다. MDPDE는 조율모수에 따라 성질이 변하게 되는데, 로버스트성과 점근효율성이 서로 상충하는 현상으로 인해 최적의 조율모수를 선택하는 것은 쉽지 않다. 본 연구에서는 MDPDE의 최적의 조율모수를 선택하는 방법으로 Fujisawa와 Eguchi (2006)가 제시한 방법과 Warwick (2006)이 제시한 방법을 소개하고, 모의실험을 통해 비교분석하였다. 연구 결과 Warwick (2006)의 방법은 특정한 경우 매우 작은 조율모수를 선택하게 될 수도 있다는 사실을 알 수 있었는데, 같은 경우에 Fujisawa와 Eguchi (2006)의 방법은 이러한 현상을 보이지 않았다. 따라서, Fujisawa와 Eguchi (2006)의 방법이 범용적으로 사용하기에 적절하다고 판단된다.

• 전기학회논문지
• /
• 제66권8호
• /
• pp.1265-1271
• /
• 2017

In this paper, the speech parameter fusion method is studied to improve the performance of the conventional emotion recognition system. For this purpose, the combination of the parameters that show the best performance by combining the cepstrum parameters and the various pitch parameters used in the conventional emotion recognition system are selected. Various pitch parameters were generated using numerical and statistical methods using pitch of speech. Performance evaluation was performed on the emotion recognition system using Gaussian mixture model(GMM) to select the pitch parameters that showed the best performance in combination with cepstrum parameters. As a parameter selection method, sequential feature selection method was used. In the experiment to distinguish the four emotions of normal, joy, sadness and angry, fifteen of the total 56 pitch parameters were selected and showed the best recognition performance when fused with cepstrum and delta cepstrum coefficients. This is a 48.9% reduction in the error of emotion recognition system using only pitch parameters.

• Nuclear Engineering and Technology
• /
• 제50권4호
• /
• pp.599-605
• /
• 2018

Fault diagnosis depends critically on the selection of sensors monitoring crucial process variables. Boolean network (BN) is composed of nodes and directed edges, where the node state is quantized to the Boolean values of True or False and is determined by the logical functions of the network parameters and the states of other nodes with edges directed to this node. Since BN can describe the fault propagation in a sensor network, it can be applied to propose sensor selection strategy for fault diagnosis. In this article, a sufficient condition for parameter identifiability of BN is first proposed, based on which the sufficient condition for fault identifiability of a sensor network is given. Then, the fault identifiability condition induces a sensor selection strategy for sensor selection. Finally, the theoretical result is applied to the fault diagnosis-oriented sensor selection for a nuclear heating reactor plant, and both the numerical computation and simulation results verify the feasibility of the newly built BN-based sensor selection strategy.