• Journal of KIISE:Software and Applications
• /
• v.36 no.5
• /
• pp.347-352
• /
• 2009

Nonnegative matrix factorization (NMF) is a popular method for multivariate analysis of nonnegative data, the goal of which is decompose a data matrix into a product of two factor matrices with all entries in factor matrices restricted to be nonnegative. NMF was shown to be useful in a task of clustering (especially document clustering). In this paper we present an algorithm for orthogonal nonnegative matrix factorization, where an orthogonality constraint is imposed on the nonnegative decomposition of a term-document matrix. We develop multiplicative updates directly from true gradient on Stiefel manifold, whereas existing algorithms consider additive orthogonality constraints. Experiments on several different document data sets show our orthogonal NMF algorithms perform better in a task of clustering, compared to the standard NMF and an existing orthogonal NMF.

• Journal of Convergence for Information Technology
• /
• v.10 no.7
• /
• pp.138-146
• /
• 2020

The paper deals with comparative study of characteristics of surrogate models and sensitivity evaluation using design of experiments in order to enhance and analysis the structure design of an automatic salt collector under various design load conditions. Orthogonal array design based on numerical analysis was used for the design of experiments. The thickness sizing variables of main structure member were considered the design factors, and the output responses were selected from the strength performances as well as the weight. The quantitative effects on responses for each design factor were evaluated from the orthogonal array experiment. Optimum design case was also identified to improve the strength performances with weight minimization. Using the orthogonal array experiment. various surrogate models such as response surface model, Kriging model, and Chebyshev orthogonal polynomial were generated. The orthogonal array experiment results were validated by the surrogate modeling results. The most suitable surrogate model was the response surface model for the exploration of design space of the automatic salt collector.

• Journal of the Korean Society of Manufacturing Process Engineers
• /
• v.20 no.5
• /
• pp.85-95
• /
• 2021

Structure design sensitivity was evaluated using the orthogonal array experimental method for passive-type deck support frame (DSF) developed for float-over installation of the offshore plant. Moreover, approximation characteristics were also reviewed based on various meta-models. The minimum weight design of the DSF is significantly important for securing both maneuvering performance and buoyancy of a ship equipped with the DSF and guaranteeing structural design safety. The performance strength of the passive type DSF was evaluated through structure analysis based on the finite element method. The thickness of main structure members was applied to design factors, and output responses were considered structure weight and strength performances. Quantitative effects on the output responses for each design factor were evaluated using the orthogonal array experimental method and analysis of variance. The optimum design case was also identified from the orthogonal array experiment results. Various meta-models, such as Chebyshev orthogonal polynomial, Kriging, response surface method, and radial basis function-based neural network, were generated from the orthogonal array experiment results. The results of the orthogonal array experiment were validated using the meta-modeling results. It was found that the radial basis function-based neural network among the meta-models could approximate the design space of the passive type DSF with the highest accuracy.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.31 no.9C
• /
• pp.889-896
• /
• 2006

The crest factor distribution of orthogonal frequency division multiplexing (OFDM) symbol sequences is evaluated and it is shown that OFDM symbol sequences with a short period are expected to have a high crest factor. The crest factor relationship between two input symbol sequences, Hamming distance D apart is also derived. Using these two results, we propose two criteria for a phase sequence set of the selected mapping (SLM) scheme and suggest the rows of the cyclic Hadamard matrix constructed from an m-sequence as the near optimal phase sequence set of the SLM scheme.

• Korean Journal of Optics and Photonics
• /
• v.6 no.3
• /
• pp.178-187
• /
• 1995

The method that determines the appropriate damping factor is studied for a lens design. When suitable damping factor is applied to the additive damped least-squares (DLS) method, the convergence and the stability of the optimization process are examined in a triplet-type photographic lens design. We calculate eigenvalues of the product of the Jacobian matrix of error functions by using the singular value decomposition (SVD) method. We adopt the median of eigenvalues as an appropriate damping factor. The convergence and the stability of the optimization process are improved by choosing the adequate damping factor for the optimization of a photographic lens. It is known that the numerical inaccuracy in the calculation of normal equation is overcome by using the orthogonal transformations of the Jacobian matrix. Therefore, a combination of the method for setting a proper damping factor and the orthogonal transformations of the Jacobian matrix is good for application to the design of an aspheric lens with high-order terms. terms.

• Transactions of the Korean Society of Mechanical Engineers A
• /
• v.28 no.6
• /
• pp.799-806
• /
• 2004

This paper presents a scheme for finding an optimal combination of design parameters affecting on the handling performance of a large truck using design of experiments. The average of the sum of peak-to-peak roll angles at the first and second part of the double lane is used as an objective function for design of experiments. Six design parameters are selected from all possible parameters affecting on the handling performance. The table of orthogonal arrays is made by 27 times simulations. A computational model of a large truck is developed by MSC/NASTRAN and MSC/ADAMS, and verified the reliability of it with the results of vehicle tests performed in a double lane change course. It is used for the simulations. Analyses of variance and factor effect of the table of orthogonal arrays are performed. This paper proposes an optimal combination of those six design parameters for improving the handling performance of the large truck.

• The Korean Journal of Applied Statistics
• /
• v.19 no.3
• /
• pp.587-598
• /
• 2006

The number of variables exceeds the number of samples in microarray data. We propose a nonparametric local linear logistic classification procedure using orthogonal components for classifying high-dimensional microarray data. The proposed method is based on the local likelihood and can be applied to multi-class classification. We applied the local linear logistic classification method using PCA, PLS, and factor analysis components as new features to Leukemia data and colon data, and compare the performance of the proposed method with the conventional statistical classification procedures. The proposed method outperforms the conventional ones for each component, and PLS has shown best performance when it is embedded in the proposed method among the three orthogonal components.

• Proceedings of the Korean Society for Technology of Plasticity Conference
• /
• 1999.03b
• /
• pp.30.1-33
• /
• 1999

Finite element simulation with experimental analysis of Taguchi's orthogonal array was carried out to know the effects of material and forming parameters on the cup earing and skewness. It was revealed that the planar anisotropy was the most influencing factor in the cup ear formation whereas blank holding force and material properties such as strength and thickness deviation at the coil edge had a relatively high effect on the cup skewness.

• Transactions of Materials Processing
• /
• v.11 no.2
• /
• pp.132-137
• /
• 2002

Using the Taguchi method, optimum process condition of lead frame blanking has been determined in the point of view of shape of blanked profile. As the main process parameters, clearance, strip holding pressure and bridge width are selected. According to the orthogonal array table, three levels of experiment have been carried out for each factor. The optimal blanking condition is analyzed with the SN ratio. It has been verified that the optimal Process condition can be determined with a combination of basic blanking experiment and experiment design method. Both the effect of each factor and gain can be judged in the quantitative manner from the analysis of SN ratio.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.10 no.4
• /
• pp.95-101
• /
• 1990

Orthogonal decomposition technique, not using normal equation, but using observation equation directly, is accepted for adjusting the geodetic network in this paper. The results of study show that the technique is the numerically stable and powerful method in network adjustment by inner constraints or weighted position parameters. Also, it is suitable to middle sized-network and is applicable to Cholesky Factor in the normal equation system.