• ETRI Journal
• /
• v.41 no.5
• /
• pp.684-695
• /
• 2019

In a cloud environment, performance degradation, or even downtime, of virtual machines (VMs) usually appears gradually along with anomalous states of VMs. To better characterize the state of a VM, all possible performance metrics are collected. For such high-dimensional datasets, this article proposes a feature extraction algorithm based on unsupervised fuzzy linear discriminant analysis with kernel (UFKLDA). By introducing the kernel method, UFKLDA can not only effectively deal with non-Gaussian datasets but also implement nonlinear feature extraction. Two sets of experiments were undertaken. In discriminability experiments, this article introduces quantitative criteria to measure discriminability among all classes of samples. The results show that UFKLDA improves discriminability compared with other popular feature extraction algorithms. In detection accuracy experiments, this article computes accuracy measures of an anomaly detection algorithm (i.e., C-SVM) on the original performance metrics and extracted features. The results show that anomaly detection with features extracted by UFKLDA improves the accuracy of detection in terms of sensitivity and specificity.

• Asian-Australasian Journal of Animal Sciences
• /
• v.21 no.10
• /
• pp.1506-1515
• /
• 2008

Two feeding trials were conducted to assess the relative bioavailability (RBV) of methionine hydroxy analogue calcium salt (MHA-Ca) in comparison to DL-methionine (DL-Met). Male Ross 308 (1-38 days) and Cobb 500 chickens (1-42 days) were used in studies 1 and 2, respectively. Experimental diets based on wheat and soybean meal or sorghum and soybean meal were fed during three phases. In both experiments graded levels of DL-Met and MHA-Ca were supplemented to Met+Cys deficient basal diets. Additionally, in experiment 1, increasing levels of a DL-Met preparation diluted with corn starch to 65% purity (DL-Met65) were supplemented. Birds were kept in floor pens and feed and water were available ad libitum. Body weights and feed consumption were recorded at the beginning and end of the experimental periods and weight gain and feed efficacy were computed subsequently. At the end of the experiments, a number of birds were slaughtered for carcass evaluation (dressing percentage, breast meat yield). Dose response data were analysed by both ANOVA and nonlinear common plateau asymptotic regression. In both experiments birds responded significantly to increasing levels of either methionine source. However, RBV of MHA-Ca compared to DL-Met was markedly (in many cases significantly) below 84%, the value which would have been expected from MHA-Ca's chemical characteristics. Excluding some extremely low RBV figures of trial 2, RBV of MHA-Ca averaged to about 63% in relation to DL-Met. In addition, supplementation of DL-Met65 allowed confirmation of nonlinear common plateau asymptotic regression to be suitable to determine RBV.

• Journal of the Korean Society for Nondestructive Testing
• /
• v.29 no.4
• /
• pp.344-350
• /
• 2009

The goal of this study is to introduce the theoretical background of acoustic nonlinearity in surface wave and to verify its characteristics by experiments. It has been known by theory that the nonlinear parameter of surface wave is proportional to the ratio of $2^{nd}$ harmonic amplitude and the power of primary component in the propagated surface wave, as like as in bulk waves. In this paper, in order to verify this characteristics we constructed a measurement system using contact angle beam transducers and measured the nonlinear parameter of surface wave in an Aluminum 6061 alloy block specimen while changing the distance of wave propagation and the input amplitude. We also considered the effect of frequency-dependent attenuation to the measurement of nonlinear parameter. Results showed good agreement with the theoretical expectation that the nonlinear parameter should be independent on the input amplitude and linearly dependent on the input amplitude and the $2^{nd}$ harmonic amplitude is linearly dependant on the propagation distance.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.16 no.10
• /
• pp.2280-2286
• /
• 2012

In this paper, we propose a novel nonlinear 3D image correlator using a fast computational integral imaging reconstruction (CIIR) method. In order to implement the fast CIIR method, the magnification process was eliminated. In the proposed correlator, elemental images of the reference and target objects are picked up by lenslet arrays. Using these elemental images, reference and target plane images are reconstructed on the output plane by means of the proposed fast CIIR method. Then, through nonlinear cross-correlations between the reconstructed reference and the target plane images, the pattern recognition can be performed from the correlation outputs. Nonlinear correlation operation can improve the recognition of 3D objects. To show the feasibility of the proposed method, some preliminary experiments are carried out and the results are presented by comparing the conventional method.

• Journal of the Institute of Electronics Engineers of Korea SC
• /
• v.48 no.5
• /
• pp.7-12
• /
• 2011

In this papwer, we propose a new feature extraction method, named as Class-augmented Kernel Principal Component Analysis (CA-KPCA), which can extract nonlinear features for classification. Among the subspace method that was being widely used for feature extraction, Class-augmented Principal Component Analysis (CA-PCA) is a recently one that can extract features for a accurate classification without computational difficulties of other methods such as Linear Discriminant Analysis (LDA). However, the features extracted by CA-PCA is still restricted to be in a linear subspace of the original data space, which limites the use of this method for various problems requiring nonlinear features. To resolve this limitation, we apply a kernel trick to develop a new version of CA-PCA to extract nonlinear features, and evaluate its performance by experiments using data sets in the UCI Machine Learning Repository.

• Journal of the Korean earth science society
• /
• v.34 no.1
• /
• pp.51-58
• /
• 2013

An appropriate determination of time step is one of the important problems in atmospheric modeling. In this study, we investigate the sensitivity of numerical solutions to time step in a nonlinear atmospheric model. For this purpose, a simple nondimensional dynamical model is employed, and numerical experiments are performed with various time steps and nonlinearity factors. Results show that numerical solutions are not sensitive to time step when the nonlinearity factor is not influentially large and truncation error is negligible. On the other hand, when the nonlinearity factor is large (i.e., in a highly nonlinear regime), numerical solutions are found to be sensitive to time step. In this situation, smaller time step increases the intensity of the spatial filter, which makes small-scale phenomena weaken. This conflicts with the fact that smaller time step generally results in more accurate numerical solutions owing to reduced truncation error. This conflict is inevitable because the spatial filter is necessary to stabilize the numerical solutions of the nonlinear model.

• Journal of the Korean Society for Aeronautical & Space Sciences
• /
• v.35 no.5
• /
• pp.375-380
• /
• 2007

The nonlinear aerodynamic analysis of wing with control surface is performed using the frequency-domain panel method. To take into consideration the nonlinear aerodynamic characteristics of wing an iterative decambering approach is introduced. The iterative decambering approach uses the known aerodynamic characteristics of airfoil to calculate the aerodynamic characteristics of wing. The multi-dimensional Newton iteration is used to account for the coupling between the different sections of wing. The present method is verified by showing that it produces results that are in good agreement with experiments. The present method will be useful for the analysis of aircraft in the conceptual design because the present method can calculate promptly the nonlinear aerodynamic characteristics of wing with a few computing resources.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.36 no.6
• /
• pp.1037-1048
• /
• 2016

In this study, nonlinear wave interaction in the presence of a uniform current is studied using numerical model, named CADMAS-SURF which is based on the Navier-Stokes equations coupled with Volume of Fluid for tracking free surface deformation. The original CADMAS-SURF developed for interaction of wave with structure is modified/extended to simulate nonlinear fluid dynamic motions within wave-current coexisting field. The capability of Numerical Wave-Current Tank (NWCT) in this study is validated by comparing with available existing laboratory experiments for both wave-following and wave-opposing current. The numerical results for interaction between wave and current are shown to be in good agreement with experimental data. Then, this study focused on the dynamic motions of the water velocity, surface elevation and vorticity within combined wave-current field in demonstrating complex nonlinear physical phenomena due to interaction between wave and current. In addition, NWCT is applied to simulate a more complex wave-current-structure field for wave propagating over a submerged breakwater associated with current. Detailed discussion including characteristics of velocity and vorticity fields and the relation between free surface and vorticity are given.

• Journal of Korean Society of Coastal and Ocean Engineers
• /
• v.19 no.5
• /
• pp.484-491
• /
• 2007

Numerical experiments using a numerical wave tank have been performed to verier the nonlinear interaction between monochromatic waves and a submerged horizontal plate. As a model for numerical wave tank, we used a higher-order Boundary Element Method(BEM) based on fully nonlinear potential flow theory and CADMAS-SURF for solving Navier Stokes equations and exact free surface conditions. Both nonlinear models are able to predict the higher harmonic generation in the shallow water region over a submerged horizontal plate. CADMAS-SURF, which involves the viscous effect, can evaluate the higher harmonic generation by flow separation and vortices at the each ends of plate. The comparison of reflection and transmission coefficients with experimental results(Patarapanich and Cheong, 1989) at different lengths and submergence depths of a horizontal plate are presented with a good agreement. It is found that the transfer of energy from the incident fundamental waves to higher harmonics becomes larger as the submergence depth ratio decreases and the length ratio increases.

• Journal of the Korean Mathematical Society
• /
• v.50 no.5
• /
• pp.1129-1163
• /
• 2013

Based on finite element discretization, two linearization approaches to the defect-correction method for the steady incompressible Navier-Stokes equations are discussed and investigated. By applying $m$ times of Newton and Picard iterations to solve an artificial viscosity stabilized nonlinear Navier-Stokes problem, respectively, and then correcting the solution by solving a linear problem, two linearized defect-correction algorithms are proposed and analyzed. Error estimates with respect to the mesh size $h$, the kinematic viscosity ${\nu}$, the stability factor ${\alpha}$ and the number of nonlinear iterations $m$ for the discrete solution are derived for the linearized one-step defect-correction algorithms. Efficient stopping criteria for the nonlinear iterations are derived. The influence of the linearizations on the accuracy of the approximate solutions are also investigated. Finally, numerical experiments on a problem with known analytical solution, the lid-driven cavity flow, and the flow over a backward-facing step are performed to verify the theoretical results and demonstrate the effectiveness of the proposed defect-correction algorithms.