• Journal of Computing Science and Engineering
• /
• 제4권2호
• /
• pp.97-109
• /
• 2010

Nonnegative matrix factorization (NMF) is a popular method for multivariate analysis of nonnegative data, which is to decompose a data matrix into a product of two factor matrices with all entries restricted to be nonnegative. NMF was shown to be useful in a task of clustering (especially document clustering), but in some cases NMF produces the results inappropriate to the clustering problems. 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. The result of orthogonal NMF can be clearly interpreted for the clustering problems, and also the performance of clustering is usually better than that of the NMF. 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.

• 한국정보과학회논문지:소프트웨어및응용
• /
• 제36권5호
• /
• pp.347-352
• /
• 2009

주어진 비음수 데이터를 두 개의 비음수 행렬의 곱의 형태로 표현하는 비음수 행렬 분해(Nonnegative Matrix Factorization)는 비음수 데이터의 다변량 분석에서 폭넓게 사용되고 있는 방법이다. 비음수 행렬 분해는 집단화(Clustering), 특히 문서의 집단화에서 유용하게 쓰일 수 있다. 본 논문에서는 주어진 문서들로부터 구성된 단어-문서 행렬을 두 개의 비음수 행렬의 곱으로 분해할 때, 그 중 하나의 행렬에 직교 제한을 주는 비음수 행렬의 직교 분해(Orthogonal Nonnegative Matrix Factorization) 방법을 다룬다. 현존하는 비음수 행렬의 직교 분해 방법은 직교 제한과 관련된 항을 더해주는 방식을 사용하지만, 여기서는 Stiefel 다양체 위에서의 실제 기울기를 직접 구하여 곱셈의 업데이트 알고리즘을 유도하였다. 다양한 문서 데이터에 대한 실험을 통해 새롭게 유도된 비음수 행렬의 직교 분해 방법이 기존의 비음수 행렬 분해나 기존의 비음수 행렬의 직교 분해보다 문서 집단화에서 우수한 성능을 나타냄을 보였다.

• 호남수학학술지
• /
• 제24권1호
• /
• pp.103-108
• /
• 2002

Applications of a column-reduced orthogonal rational matrix function to McMillan degrees and Wiener-Hopf factorizations are considered.

• 대한수학회보
• /
• 제39권3호
• /
• pp.359-376
• /
• 2002

We give a new interpretation of Darboux transforms in the context of orthogonal polynomials and find conditions in or-der for any Darboux transform to yield a new set of orthogonal polynomials. We also discuss connections between Darboux trans-forms and factorization of linear differential operators which have orthogonal polynomial eigenfunctions.

• 대한수학회지
• /
• 제45권6호
• /
• pp.1613-1622
• /
• 2008

Let G be a graph with vertex set V(G) and edge set E(G), and let g, f be two nonnegative integer-valued functions defined on V(G) such that $g(x)\;{\leq}\;f(x)$ for every vertex x of V(G). We use $d_G(x)$ to denote the degree of a vertex x of G. A (g, f)-factor of G is a spanning subgraph F of G such that $g(x)\;{\leq}\;d_F(x)\;{\leq}\;f(x)$ for every vertex x of V(F). In particular, G is called a (g, f)-graph if G itself is a (g, f)-factor. A (g, f)-factorization of G is a partition of E(G) into edge-disjoint (g, f)-factors. Let F = {$F_1$, $F_2$, ..., $F_m$} be a factorization of G and H be a subgraph of G with mr edges. If $F_i$, $1\;{\leq}\;i\;{\leq}\;m$, has exactly r edges in common with H, we say that F is r-orthogonal to H. If for any partition {$A_1$, $A_2$, ..., $A_m$} of E(H) with $|A_i|=r$ there is a (g, f)-factorization F = {$F_1$, $F_2$, ..., $F_m$} of G such that $A_i\;{\subseteq}E(F_i)$, $1\;{\leq}\;i\;{\leq}\;m$, then we say that G has (g, f)-factorizations randomly r-orthogonal to H. In this paper it is proved that every (0, mf - (m - 1)r)-graph has (0, f)-factorizations randomly r-orthogonal to any given subgraph with mr edges if $f(x)\;{\geq}\;3r\;-\;1$ for any $x\;{\in}\;V(G)$.

• 경영과학
• /
• 제21권2호
• /
• pp.43-60
• /
• 2004

Every iteration of interior-point methods of large scale optimization requires computing at least one orthogonal projection. In the practice, symmetric variants of the Gaussian elimination such as Cholesky factorization are accepted as the most efficient and sufficiently stable method. In this paper several specific implementation issues of the symmetric factorization that can be applied for solving such equations are discussed. The code called McSML being the result of this work is shown to produce comparably sparse factors as another implementations in the $MATLAB^{***}$ environment. It has been used for computing projections in an efficient implementation of self-regular based interior-point methods, McIPM. Although primary aim of developing McSML was to embed it into an interior-point methods optimizer, the code may equally well be used to solve general large sparse systems arising in different applications.

• 대한수학회지
• /
• 제46권2호
• /
• pp.281-294
• /
• 2009

Accuracy of the scaling function is very crucial in wavelet theory, or correspondingly, in the study of wavelet filter banks. We are mainly interested in vector-valued filter banks having matrix factorization and indicate how to choose block central symmetric matrices to construct multi-wavelets with suitable accuracy.

• Communications for Statistical Applications and Methods
• /
• 제31권5호
• /
• pp.521-533
• /
• 2024

As the concentration of fine dust has recently increased, numerous related studies are being conducted to address this issue. Aerosol optical depth (AOD) is a vital atmospheric parameter for measuring the optical properties of aerosols in the atmosphere, providing crucial information related to fine dust. In this paper, we apply three dimension reduction methods, nonnegative matrix factorization (NMF), empirical orthogonal functions (EOF) analysis and independent component analysis (ICA), to AOD data to analyze the patterns of fine dust in the East Asia region. Through a comparison of three dimension reduction methods, we observe that some patterns are observed in all three method, while some information are only extracted in a specific method. Additionally, we forecast AOD levels based on three methods, and compare the predictive performance of the three methodologies.

• Journal of Communications and Networks
• /
• 제5권4호
• /
• pp.328-343
• /
• 2003

Orthogonal pulse-shape modulation using Hermite pulses for ultra-wideband communications is reviewed. Closedform expressions of cross-correlations among Hermite pulses and their corresponding transmit and receive waveforms are provided. These show that the pulses lose orthogonality at the receiver in the presence of differentiating antennas. Using these expressions, an algebraic model is established based on the projections of distorted receive waveforms onto the orthonormal basis given by the set of normalized orthogonal Hermite pulses. Using this new matrix model, a number of pulse-shape modulation schemes are analyzed and a novel orthogonal design is proposed. In the proposed orthogonal design, transmit waveforms are constructed as combinations of elementary Hermites with weighting coefficients derived by employing the Gram-Schmidt (QR) factorization of the differentiating distortion model’s matrix. The design ensures orthogonality of the vectors at the output of the receiver bank of correlators, without requiring compensation for the distortion introduced by the antennas. In addition, a new set of elementary Hermite Pulses is proposed which further enhances the performance of the new design while enabling a simplified hardware implementation.

• Journal of applied mathematics & informatics
• /
• 제29권3_4호
• /
• pp.1037-1042
• /
• 2011

Multiwavelet coefficients can be constructed from the multi-scaling coefficients by using the factorization for paraunitary matrices. In this paper we present a procedure for parametrizing all possible multi-wavelet coefficients corresponding to the multiscaling coefficients of DGHM type.