• Journal of the Korean Data and Information Science Society
• /
• v.22 no.5
• /
• pp.999-1005
• /
• 2011

This paper proposes a weighted least squares support vector machine for asymmetric least squares regression. This method achieves nonlinear prediction power, while making no assumption on the underlying probability distributions. The cross validation function is introduced to choose optimal hyperparameters in the procedure. Experimental results are then presented which indicate the performance of the proposed model.

• Journal of applied mathematics & informatics
• /
• v.34 no.1_2
• /
• pp.19-34
• /
• 2016

In this paper, we present a split least-squares characteristic mixed finite element method(MFEM) to get the approximate solutions of the convection dominated Sobolev equations. First, to manage both convection term and time derivative term efficiently, we apply a least-squares characteristic MFEM to get the system of equations in the primal unknown and the flux unknown. Then, we obtain a split least-squares characteristic MFEM to convert the coupled system in two unknowns derived from the least-squares characteristic MFEM into two uncoupled systems in the unknowns. We theoretically prove that the approximations constructed by the split least-squares characteristic MFEM converge with the optimal order in L² and H¹ normed spaces for the primal unknown and with the optimal order in L² normed space for the flux unknown. And we provide some numerical results to confirm the validity of our theoretical results.

• Journal of the Korean Mathematical Society
• /
• v.54 no.5
• /
• pp.1557-1571
• /
• 2017

The L-curve method is a parametric plot of interrelation between the residual norm of the least squares problem and the solution norm. However, the L-curve method may be hard to apply to the total least squares problem due to its no closed form solution of the regularized total least squares problems. Thus the sequence of the solution norm under the fixed regularization parameter and its corresponding residual need to be found with an efficient manner. In this paper, we suggest an efficient algorithm to find the sequence of the solutions and its residual in order to plot the L-curve for the total least squares problems. In the numerical experiments, we present that the proposed algorithm successfully and efficiently plots fairly 'L' like shape for some practical regularized total least squares problems.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.25 no.10
• /
• pp.1605-1612
• /
• 2001

Least-squares meshfree method is presented. Conventional meshfree methods based on the Galerkin formulation suffer from inaccurate numerical integration. Least-squares formulation exhibits rather different integration-related characteristics. It is demonstrated through numerical examples that least-squares formulation is much more robust to integration errors than the Galerkin's. Therefore efficient meshfree methods can be devised by combining very simple integration algorithms and least-squares formulation.

• Journal of applied mathematics & informatics
• /
• v.4 no.2
• /
• pp.501-512
• /
• 1997

In multiple linear regression model we have presupposed assumptions (independence normality variance homogeneity and so on) on error term. When case weights are given because of variance heterogeneity we can estimate efficiently regression parameter using weighted least squares estimator. Unfortunately this estimator is sen-sitive to outliers like ordinary least squares estimator. Thus in this paper we proposed some statistics for detection of outliers in weighted least squares regression.

• Journal of the Korean Statistical Society
• /
• v.28 no.2
• /
• pp.151-165
• /
• 1999

This paper considers the trimmed least squares estimator of the autoregression parameter in the unstable AR(1) model: X\ulcorner=ØX\ulcorner+$\varepsilon$\ulcorner, where $\varepsilon$\ulcorner are iid random variables with mean 0 and variance $\sigma$$^2$> 0, and Ø is the real number with │Ø│=1. The trimmed least squares estimator for Ø is defined in analogy of that of Welsh(1987). The limiting distribution of the trimmed least squares estimator is derived under certain regularity conditions.

• Communications for Statistical Applications and Methods
• /
• v.13 no.2
• /
• pp.233-241
• /
• 2006

Classical principal component analysis (PCA) can be formulated as finding the linear subspace that best accommodates multidimensional data points in the sense that the sum of squared residual distances is minimized. As alternatives to such LS (least squares) fitting approach, we produce LMS (least median of squares) and LTS (least trimmed squares)-type PCA by minimizing the median of squared residual distances and the trimmed sum of squares, in a similar fashion to Rousseeuw (1984)'s alternative approaches to LS linear regression. Proposed methods adopt the data-driven optimization algorithm of Croux and Ruiz-Gazen (1996, 2005) that is conceptually simple and computationally practical. Numerical examples are given.

• The Journal of the Acoustical Society of Korea
• /
• v.29 no.6
• /
• pp.374-380
• /
• 2010

It is known that the problem of FIR filtering with noisy input and output data can be solved by a total least squares (TLS) estimation. It is also known that the performance of the TLS estimation is very sensitive to the ratio between the variances of the input and output noises. In this paper, we propose a convex combination algorithm between the ordinary recursive LS based TLS (RTLS) and the ordinary recursive LS (RLS). This combined algorithm is robust to the noise variance ratio and has almost the same complexity as the RTLS. Simulation results show that the proposed algorithm performs near TLS in noise variance ratio ${\gamma}{\approx}1$ and that it outperforms TLS and LS in the rage of 2 < $\gamma$ < 20. Consequently, the practical workability of the TLS method applied to noisy data has been significantly broadened.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.28 no.12
• /
• pp.2019-2031
• /
• 2004

The least-squares formulation for rigid-plasticity based on J$_2$-flow rule and infinitesimal theory and its meshfree implementation using moving least-squares approximation are proposed. In the least-squares formulation the squared residuals of the constitutive and equilibrium equations are minimized. Those residuals are represented in a form of first-order differential system using the velocity and stress components as independent variables. For the enforcement of the boundary and frictional contact conditions, penalty scheme is employed. Also the reshaping of nodal supports is introduced to avoid the difficulties due to the severe local deformation near the contact interface. The proposed least-squares meshfree method does not require any structure of extrinsic cells during the whole process of analysis. Through some numerical examples of metal forming processes, the validity and effectiveness of the method are investigated.

• Journal of the Korean Mathematical Society
• /
• v.52 no.2
• /
• pp.349-372
• /
• 2015

This paper concerns with exploiting an oblique projection technique to solve a general class of large and sparse least squares problem over symmetric arrowhead matrices. As a matter of fact, we develop the conjugate gradient least squares (CGLS) algorithm to obtain the minimum norm symmetric arrowhead least squares solution of the general coupled matrix equations. Furthermore, an approach is offered for computing the optimal approximate symmetric arrowhead solution of the mentioned least squares problem corresponding to a given arbitrary matrix group. In addition, the minimization property of the proposed algorithm is established by utilizing the feature of approximate solutions derived by the projection method. Finally, some numerical experiments are examined which reveal the applicability and feasibility of the handled algorithm.