• Journal of the Korean Data and Information Science Society
• /
• v.18 no.3
• /
• pp.735-744
• /
• 2007

Support vector machine(SVM) is capable of providing a more complete description of the linear and nonlinear relationships among random variables. In this paper we propose a sparse kernel regression(SKR) to overcome a weak point of SVM, which is, the steep growth of the number of support vectors with increasing the number of training data. The iterative reweighted least squares(IRWLS) procedure is used to solve the optimal problem of SKR with a Laplacian prior. Furthermore, the generalized cross validation(GCV) function is introduced to select the hyper-parameters which affect the performance of SKR. Experimental results are then presented which illustrate the performance of the proposed procedure.

• Communications for Statistical Applications and Methods
• /
• v.15 no.3
• /
• pp.441-450
• /
• 2008

In this paper we propose a new method for solving multiclass problem with least squares support vector machine(LS-SVM) regression. This method implements one-against-all scheme which is as accurate as any other approach. We also propose cross validation(CV) method to select effectively the optimal values of hyper-parameters which affect the performance of the proposed multiclass method. Experimental results are then presented which indicate the performance of the proposed multiclass method.

• Journal of Information Processing Systems
• /
• v.15 no.1
• /
• pp.116-126
• /
• 2019

In this paper, an improved one-against-one support vector machine algorithm is used to classify multiple power quality disturbances. To solve the problem of parameter selection, an improved particle swarm optimization algorithm is proposed to optimize the parameters of the support vector machine. By proposing a new inertia weight expression, the particle swarm optimization algorithm can effectively conduct a global search at the outset and effectively search locally later in a study, which improves the overall classification accuracy. The experimental results show that the improved particle swarm optimization method is more accurate than a grid search algorithm optimization and other improved particle swarm optimizations with regard to its classification of multiple power quality disturbances. Furthermore, the number of support vectors is reduced.

• Journal of Electrical Engineering and Technology
• /
• v.12 no.2
• /
• pp.830-839
• /
• 2017

Support vector machine (SVM) is introduced as an effective fault diagnosis technique based on dissolved gases analysis (DGA) for oil-immersed transformers with maximum generalization ability; however, the applicability of the SVM is highly affected due to the difficulty of selecting the SVM parameters appropriately. Therefore, a novel approach combing SVM with improved imperialist competitive algorithm (IICA) for fault diagnosis of oil-immersed transformers was proposed in the paper. The improved ICA, which is proved to be an effective optimization approach, is employed to optimize the parameters of SVM. Cross validation and normalizations were applied in the training processes of SVM and the trained SVM model with the optimized parameters was established for fault diagnosis of oil-immersed transformers. Three classification benchmark sets were studied based on particle swarm optimization SVM (PSOSVM) and IICASVM with four multiple classification schemes to select the best scheme for transformer fault diagnosis. The results show that the proposed model can obtain higher diagnosis accuracy than other methods. The comparisons confirm that the proposed model is an effective approach for classification problems.

• Physical Therapy Korea
• /
• v.21 no.3
• /
• pp.80-88
• /
• 2014

Previous studies have reported that decreased cognitive ability has been consistently associated with significant declines in performance of one or both tasks under a dual-task walking condition. This study examined the relationship between specific cognitive abilities and the dual-task costs (DTCs) of spatio-temporal gait parameters in stroke patients. The spatio-temporal gait parameters were measured among 30 stroke patients while walking with and without a cognitive task (Stroop Word-Color Task) at the study participant's preferred walking speed. Cognitive abilities were measured using Computerized Neuropsychological Testing. Pearson's correlation coefficients (r) were calculated to quantify the associations between the neuropsychological measures and the DTCs in the spatio-temporal gait parameters. Moderate to strong correlations were found between the Auditory Continuous Performance test (ACPT) and the DTCs of the Single Support Time of Non-paretic (r=.37), the Trail Making A (TMA) test and the DTCs of Velocity (r=.71), TMA test and the DTCs of the Step Length of Paretic (r=.37), TMA test and the DTCs of the Step Length Non-paretic (r=.36), the Trail Making B (TMB) test and the DTCs of Velocity (r=.70), the Stroop Word-Color test and the DTCs of Velocity (r=-.40), Visual-span Backward (V-span B) test and the DTCs of Velocity (r=-.41), V-span B test and the DTCs of the Double Support Time of Non-paretic (r=.38), Digit-span Forward test and the DTCs of the Step Time of Non-paretic (r=-.39), and Digit-span Backward test and the DTCs of the Single Support Time of Paretic (r=.36). Especially TMA test and TMB test were found to be more strongly correlated to the DTCs of gait velocity than the other correlations. Understanding these cognitive features will provide guidance for identifying dual- task walking ability.

• Journal of The Korean Astronomical Society
• /
• v.45 no.2
• /
• pp.31-38
• /
• 2012

In this study we apply Support Vector Machine (SVM) to the prediction of geo-effective halo coronal mass ejections (CMEs). The SVM, which is one of machine learning algorithms, is used for the purpose of classification and regression analysis. We use halo and partial halo CMEs from January 1996 to April 2010 in the SOHO/LASCO CME Catalog for training and prediction. And we also use their associated X-ray flare classes to identify front-side halo CMEs (stronger than B1 class), and the Dst index to determine geo-effective halo CMEs (stronger than -50 nT). The combinations of the speed and the angular width of CMEs, and their associated X-ray classes are used for input features of the SVM. We make an attempt to find the best model by using cross-validation which is processed by changing kernel functions of the SVM and their parameters. As a result we obtain statistical parameters for the best model by using the speed of CME and its associated X-ray flare class as input features of the SVM: Accuracy=0.66, PODy=0.76, PODn=0.49, FAR=0.72, Bias=1.06, CSI=0.59, TSS=0.25. The performance of the statistical parameters by applying the SVM is much better than those from the simple classifications based on constant classifiers.

• Korean Journal of Applied Biomechanics
• /
• v.15 no.4
• /
• pp.13-24
• /
• 2005

The purpose of this study was to compare kinematic data between experts and novices, and identify difference kinematic parameters changing direction to kick in penalty kick of soccer play. Novice subjects were 5 high school students Who has never been experienced a soccer player, and expert subjects were 5 competitive high school soccer players. The 3-d angle was calculated by Euler's Angle by inertial axis and local axis with three-dimensional cinematography. Kinematic parameters in this study consisted of angles of knee joints, hip joints, lower trunk and upper trunk when the support foot was contacted on ground and kicking foot impacted the ball. The difference of angle of knee joints in the flexion/extension was insignificantly showed below $4{\sim}9^{\circ}$ in groups and directions of ball at the time of support and impact. But the difference of angle of hip joint was significant in groups and directions of ball at the time of support and impact. Specially the right hip joint of experts were more flexed about $12^{\circ}$($43.99{\pm}6.17^{\circ}$ at left side, $31.87{\pm}4.49^{\circ}$ at right side), less abducted about $10^{\circ}$ ($-31.27{\pm}4.49^{\circ}$ at left side, $-41.97{\pm}6.67^{\circ}$ at right side) at impact when they kicked a ball to the left side of goalpost. The difference of amplitude angle in the trunk was significantly shown at upper trunk not lower trunk. The upper trunk was external rotated about $30^{\circ}$ (novice' angle was $-16.3{\pm}17.08^{\circ}$, expert's angle was $-43.73{\pm}12.79^{\circ}$) at impact. Therefore the significant difference of kinematic characteristics could be found at the right hip joint and the upper trunk at penalty kick depending on the direction of kicking.

• Journal of Korean Port Research
• /
• v.15 no.1
• /
• pp.37-46
• /
• 2001

Fuzzy logic control(FLC) is composed of three parts : fuzzy rule-bases, membership functions, and scaling factors. Well-defined fuzzy rule-base should contain proper physical intuition on the plant, so are needed lots of experiences of the skillful expert. When membership functions are considered, some parameters on the memberships function such as function shape, support, allocation density should be selected well. The rule of scaling factors is 'scaling'(amplifying or reducing) for both input and output signals of the FLC to fit in the membership function support and to operate the plant intentionally. To get a better performance of the FLC, it is necessary to adjust the parameters of the FLC. In general, the adaptation of the scaling factors is the most effective adjustment scheme, compared with that of the fuzzy rule-base or membership function parameters. This study proposes the adaptation scheme of the scaling factors. When the adaptation is performed on-line, the stability of the adaptive FLC should be guaranteed. The stable FLC system can be designed with stability analysis in the sense of Lyapunov stability. To adapt the scaling factors for the error signals, the concept of the conventional MRAC would be introduced into slightly modified form. A tracking accuracy of the control system would be enhanced by the modified shape and support of the membership function. The simulation is achieved on the pilot plant with the hydraulic steering control of a UCT(Unmanned Container Transporter) of which modeling dynamics have lots of severe uncertainties and modeling errors.

• Journal of the Korean Data and Information Science Society
• /
• v.20 no.4
• /
• pp.749-755
• /
• 2009

Support vector classification (SVC) provides more complete description of the lin-ear and nonlinear relationships between input vectors and classifiers. In this paper. we propose the sparse kernel classifier to solve the optimization problem of classification with a modified hinge loss function and absolute loss function, which provides the efficient computation and the sparsity. We also introduce the generalized cross validation function to select the hyper-parameters which affects the classification performance of the proposed method. Experimental results are then presented which illustrate the performance of the proposed procedure for classification.

• Journal of the Korean Data and Information Science Society
• /
• v.14 no.2
• /
• pp.337-343
• /
• 2003

In this paper we present the prediction interval estimation method using bootstrap method for least squares support vector machine(LS-SVM) regression, which allows us to perform even nonlinear regression by constructing a linear regression function in a high dimensional feature space. The bootstrap method is applied to generate the bootstrap sample for estimation of the covariance of the regression parameters consisting of the optimal bias and Lagrange multipliers. Experimental results are then presented which indicate the performance of this algorithm.