• Journal of the Korean Data and Information Science Society
• /
• 제23권4호
• /
• pp.765-775
• /
• 2012

교차타당성은 커널추정량의 평활모수인 띠폭의 선택 방법으로 흔히 활용되고 있다. 연속인 확률밀도함수의 커널추정량의 띠폭 선택으로 널리 쓰이는 교차타당성 방법으로는 최대가능도교차타당성과 더불어 최소제곱교차타당성과 편의교차타당성이 있다. 확률밀도함수가 하나의 불연속점을 가질 때, Huh (2012)는 불연속점 추정을 위한 커널추정량의 띠폭 선택으로 최대가능도교차타당성을 이용한 방법을 제시하였다. 본 연구에서는 Huh (2012)에 의해 최대가능도교차타당성으로 제안된 띠폭선택의 방법과 같이 한쪽방향커널함수를 이용한 최소제곱교차타당성과 편의교차타당성으로 띠폭 선택 방법을 제시하고, 이들 띠폭 선택 방법들과 Huh (2012)의 최대가능도교차타당성을 이용한 띠폭 선택 방법을 모의실험을 통하여 비교연구 하고자 한다.

• Journal of the Korean Data and Information Science Society
• /
• 제28권5호
• /
• pp.1153-1165
• /
• 2017

구조방정식모형은 사회과학 및 행동과학 연구 분야에서 이론검정을 위해 주로 사용되는 통계방법이다. 최근 이 통계기법에 대한 방법론적 이슈로서 모형명세화 오류와 소표본 문제가 부각되고 있다. 그런데 이 문제들이 구조방정식모형의 대표 추정 방법인 최대우도법에 위한 이론검정에 어떤 영향을 주는지에 대해 여전히 명확하지 않다. 따라서 본 연구에서 최대우도법 그러고 이에 대한 대안으로 개발된 2단계최소자승법과 2단계능형최소자승법을 정확도와 검정력 관점에서 시뮬레이션을 통해 체계적으로 비교해 본다. 이 실험 결과에 따르면, 모형이 정확하게 설정된 경우, 정확도 기준에서 추정방법들 간의 차이는 미미했다. 하지만 모형오류가 발생한 경우, 2단계능형최소자승법은 다른 방법들보다 표본 크기가 작을 때 훨씬 더 정확한 모수추정치를 산출해 내었다. 그러고 이 방법은 명세화 오류에 관계없이 표본 크기가 작을 때에도 제 2종 오류 (Type II error) 수준이 상대적으로 작거나 만족할만한 수준의 검정력을 보여주었다. 이에 반해 다른 두 방법들은 표본이 작은 경우 또는 명세화 오류가 있는 경우 상당히 높은 수준의 제 2종 오류를 나타내었다.

• Archives of Pharmacal Research
• /
• 제10권1호
• /
• pp.1-8
• /
• 1987

One of the problems of derivatie spectrophotometry, the decrease of signal-to-noise ratio by derivative operations, was solved by three concepts of digital filtering, ensemble averaging, least squares polynomial smoothing and Fourier smoothing. The suthors made several compouter programs written in APPLE SOFT BASIC language for the actual applications of the concepts of these digital filters on UV spectrophotometer system. As a result, ensemble averaging could not be used as a routine operation for the spectrophotometer used. The maximum S/N ratio enhancement factors achieved by least squares polynomial smoothing were 6.17 and 7.47 for the spectra of Gaussian and Lorentzian distribution models, and by Fourier smoothing 16.42 and 11.78 for the spectra of two models, respectively.

• 한국컴퓨터그래픽스학회논문지
• /
• 제2권2호
• /
• pp.11-20
• /
• 1996

산포된 자료 보간을 응용하는 분야에는 모델링, 자연현상 가시화 등을 비롯하여 여러 가지를 들 수 있다. 사면체 분할은 사차원적 공간 형성을 위한 전 처리 단계 중의 하나이다. 본 논문은 다양한 사면체 분할법인, Delaunay, least squares fitting, gradient difference, 와 jump in normal direction derivatives 들을 논의하였다. 본 논문은 사면체 영역을 가시화 함으로써, 사면체 분할법들을 구별시키고, 사면체 영역을 바탕으로 보간된 공간상의 등사치를 수치적 뿐만 아니라 시각적으로 가시화 하여 그 정확도를 비교 분석할 수 있는 방법을 제시하였다.

• Communications for Statistical Applications and Methods
• /
• 제10권2호
• /
• pp.595-601
• /
• 2003

In this paper we propose an on line training method for classification based on least squares support vector machine. Proposed method enables the computation cost to be reduced and the training to be peformed incrementally, With the incremental formulation of an inverse matrix in optimization problem, current information and new input data can be used for building the new inverse matrix for the estimation of the optimal bias and Lagrange multipliers, so the large scale matrix inversion operation can be avoided. Numerical examples are included which indicate the performance of proposed algorithm.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제4권1호
• /
• pp.45-61
• /
• 2010

In the communication channel, the received signal is affected by many factors that can cause errors. These effects mean that received signal strength (RSS) based methods incur more errors in measuring distance and consequently result in low precision in the location detection process. As one of the approaches to overcome these problems, we propose using direction calibration to improve the performance of the RSS-based method for distance measurement, and sequentially a weighted least squares (WLS) method using reliability factors in conjunction with a conventional RSS weighting matrix is proposed to solve an over-determined localization process. The proposed scheme focuses on the features of the RSS method to improve the performance, and these effects are proved by the simulation results.

• Communications for Statistical Applications and Methods
• /
• 제24권3호
• /
• pp.317-324
• /
• 2017

An influence measure known as Cook's distance has been used for judging the influence of each observation on the least squares estimate of the parameter vector. The distance does not reflect the distributional property of the change in the least squares estimator of the regression coefficients due to case deletions: the distribution has a covariance matrix of rank one and thus it has a support set determined by a line in the multidimensional Euclidean space. As a result, the use of Cook's distance may fail to correctly provide information about influential observations, and we study some reasons for the failure. Three illustrative examples will be provided, in which the use of Cook's distance fails to give the right information about influential observations or it provides the right information about the most influential observation. We will seek some reasons for the wrong or right provision of information.

• 대한의용생체공학회:의공학회지
• /
• 제26권2호
• /
• pp.87-93
• /
• 2005

A new method for real-time estimating left ventricular relaxation time constant (T) from the left ventricular (LV) pressure waveform, based on the isovolumic relaxation model, is proposed. The presented method uses a recursive least squares (RLS) algorithm to accomplish real-time estimation. A new criterion to detect the end-point of the isovolumic relaxation period (IRP) for the estimation of T is also introduced, which is based on the pattern analysis of mean square errors between the original and reconstructed pressure waveforms. We have verified the performance of the new method in over 4,600 beats obtained from 70 patients. The results demonstrate that the proposed method provides more stable and reliable estimation of τ than the conventional 'off-line' methods.

• Journal of the Korean Data and Information Science Society
• /
• 제27권3호
• /
• pp.827-833
• /
• 2016

In this paper, we propose a deep least squares support vector machine (LS-SVM) for regression problems, which consists of the input layer and the hidden layer. In the hidden layer, LS-SVMs are trained with the original input variables and the perturbed responses. For the final output, the main LS-SVM is trained with the outputs from LS-SVMs of the hidden layer as input variables and the original responses. In contrast to the multilayer neural network (MNN), LS-SVMs in the deep LS-SVM are trained to minimize the penalized objective function. Thus, the learning dynamics of the deep LS-SVM are entirely different from MNN in which all weights and biases are trained to minimize one final error function. When compared to MNN approaches, the deep LS-SVM does not make use of any combination weights, but trains all LS-SVMs in the architecture. Experimental results from real datasets illustrate that the deep LS-SVM significantly outperforms state of the art machine learning methods on regression problems.

• Journal of the Korean Data and Information Science Society
• /
• 제21권3호
• /
• pp.561-567
• /
• 2010

Multiclassification is typically performed using voting scheme methods based on combining a set of binary classifications. In this paper we use multiclassification method with a hat matrix of least squares support vector machine (LS-SVM), which can be regarded as the revised one-against-all method. To tackle multiclass problems for large data, we use the $Nystr\ddot{o}m$ approximation and the quadratic Renyi entropy with estimation in the primal space such as used in xed size LS-SVM. For the selection of hyperparameters, generalized cross validation techniques are employed. Experimental results are then presented to indicate the performance of the proposed procedure.