• Title/Summary/Keyword: Singular value decomposition

Search Result 371, Processing Time 0.027 seconds

KOREAN TOPIC MODELING USING MATRIX DECOMPOSITION

  • June-Ho Lee;Hyun-Min Kim
    • East Asian mathematical journal
    • /
    • v.40 no.3
    • /
    • pp.307-318
    • /
    • 2024
  • This paper explores the application of matrix factorization, specifically CUR decomposition, in the clustering of Korean language documents by topic. It addresses the unique challenges of Natural Language Processing (NLP) in dealing with the Korean language's distinctive features, such as agglutinative words and morphological ambiguity. The study compares the effectiveness of Latent Semantic Analysis (LSA) using CUR decomposition with the classical Singular Value Decomposition (SVD) method in the context of Korean text. Experiments are conducted using Korean Wikipedia documents and newspaper data, providing insight into the accuracy and efficiency of these techniques. The findings demonstrate the potential of CUR decomposition to improve the accuracy of document clustering in Korean, offering a valuable approach to text mining and information retrieval in agglutinative languages.

An Orthogonal Representation of Estimable Functions

  • Yi, Seong-Baek
    • Communications for Statistical Applications and Methods
    • /
    • v.15 no.6
    • /
    • pp.837-842
    • /
    • 2008
  • Students taking linear model courses have difficulty in determining which parametric functions are estimable when the design matrix of a linear model is rank deficient. In this note a special form of estimable functions is presented with a linear combination of some orthogonal estimable functions. Here, the orthogonality means the least squares estimators of the estimable functions are uncorrelated and have the same variance. The number of the orthogonal estimable functions composing the special form is equal to the rank of the design matrix. The orthogonal estimable functions can be easily obtained through the singular value decomposition of the design matrix.

New Upper Bounds for the CALE: A Singular Value Decomposition Approach

  • Savov, Svetoslav G.;Popchev, Ivan P.
    • International Journal of Control, Automation, and Systems
    • /
    • v.6 no.2
    • /
    • pp.288-294
    • /
    • 2008
  • Motivated by the fact that upper solution bounds for the continuous Lyapunov equation are valid under some very restrictive conditions, an attempt is made to extend the set of Hurwitz matrices for which such bounds are applicable. It is shown that the matrix set for which solution bounds are available is only a subset of another stable matrices set. This helps to loosen the validity restriction. The new bounds are illustrated by examples.

Reverberation Characterization and Suppression by Means of Low Rank Approximation (낮은 계수 근사법을 이용한 표준 잔향음 신호 획득 및 제거 기법)

  • 윤관섭;최지웅;나정열
    • The Journal of the Acoustical Society of Korea
    • /
    • v.21 no.5
    • /
    • pp.494-502
    • /
    • 2002
  • In this paper, the Low Rank Approximation (LRA) method to suppress the interference of signals from temporal fluctuations is applied. The reverberation signals and temporally fluctuating signals are separated from the measured data using the Ink. The Singular value decomposition (SVD) method is applied to extract the low rank and the temporally stable reverberation was extracted using the LRA. The reverberation suppression is performed on the LRA residual value obtained by removing the approximate reverberation signals. In overall, the method can be applied to the suppression of reververation in active sonar system as well as to the modeling of reverberation.

NMR Solvent Peak Suppression by Piecewise Polynomial Truncated Singular Value Decomposition Methods

  • Kim, Dae-Sung;Lee, Hye-Kyoung;Won, Young-Do;Kim, Dai-Gyoung;Lee, Young-Woo;Won, Ho-Shik
    • Bulletin of the Korean Chemical Society
    • /
    • v.24 no.7
    • /
    • pp.967-970
    • /
    • 2003
  • A new modified singular value decomposition method, piecewise polynomial truncated SVD (PPTSVD), which was originally developed to identify discontinuity of the earth's radial density function, has been used for large solvent peak suppression and noise elimination in nuclear magnetic resonance (NMR) signal processing. PPTSVD consists of two algorithms of truncated SVD (TSVD) and L₁ problems. In TSVD, some unwanted large solvent peaks and noise are suppressed with a certain soft threshold value, whereas signal and noise in raw data are resolved and eliminated in L₁ problems. These two algorithms were systematically programmed to produce high quality of NMR spectra, including a better solvent peak suppression with good spectral line shapes and better noise suppression with a higher signal to noise ratio value up to 27% spectral enhancement, which is applicable to multidimensional NMR data processing.

A Research for Appling Singular Value Decomposition to Collaborative Filtering for Coping With the Sparsity of Rating matrix (협력적 여과에서 평가 행렬의 희소성 문제를 해결하기 위한 Singular Value Decomposition의 적용 방법에 관한 연구)

  • Jeong, Jun;Jeong, Dae-jin;Kim, Yong-Han;Rhee, Phill-Kyu
    • Proceedings of the Korea Inteligent Information System Society Conference
    • /
    • 2000.04a
    • /
    • pp.317-322
    • /
    • 2000
  • 인터넷의 발달로 사용자들은 인터넷에서 필요한 정보를 습득할 수 있을 뿐만 아니라, 생활에 필요한 여러 가지 활동들을 할 수 있게 되었다. 특히 주목받는 부분은 구매 활동이다. 따라서 수많은 기업들이 사람들의 구매 활동에 관련된 전자상거래에 투자하고 있고, 현재 Amazon.com 등과 같은 세계적인 사이트들이 서비스를 실시하고 있다. 또한, 전자상거래 사이트들은 사용자들의 구매 활동을 도와주기 위해 추천 시스템의 도입을 추진하고 있다. 추천 시스템은 사용자들로부터 얻어진 정보를 학습하여 이용 가능한 상품 중에서 고객이 좋아할 만한 것은 추천해 주는 시스템이다. 본 논문에서는 추천 시스템에서 사용되는 주요한 방법인 협력적 여과방법에서 초기 rating 행렬의 희소성 문제를 해결하기 위하여 Singular Value decompositon의 적용 방법을 제안하고 있다.

  • PDF

Digital Watermarking based on Wavelet Transform and Singular Value Decomposition(SVD) (웨이블릿 변환과 특이치 분해에 기반한 디지털 워터마킹)

  • 김철기;차의영
    • The Journal of Korean Institute of Communications and Information Sciences
    • /
    • v.27 no.6A
    • /
    • pp.602-609
    • /
    • 2002
  • In this paper, we propose an robust invisible watermarking method using wavelet transform and singular value decomposition for the ownership protection. of images. For this method, after we decompose the original image in three level using wavelet transform, we use singular value decomposition based key depended watermark insertion method in the lowest band $LL_3.$ And we also watermark using DCT for extraction of watermark and verification of robustness. In the experiments, we found that it had a good quality and robustness in attack such as compression, image processing, geometric transformation and noises. Especially, we know that this method have very high extraction ratio against nose and JPEG compression. And Digimarc's method can not extract watermark in 80 percent compression ratio of JPEG, but the proposed method can extract well.

Updating Algorithms of Finite Element Model Using Singular Value Decomposition and Eigenanalysis (특이값 분해와 고유치해석을 이용한 유한요소모델의 개선)

  • 김홍준;박영필
    • Journal of KSNVE
    • /
    • v.9 no.1
    • /
    • pp.163-173
    • /
    • 1999
  • Precise and reasonable modelling is necessary and indispensable to the analysis of dynamic characteristics of mechanical structures. Also. the effective prediction of the change of modal properties due to the variation of design parameters is required especially for the application of finite element method to the structural dynamics problems. To meet those necessity and requirement, three model updating algorithms are proposed for finite element methods. Those algorithms are based on sensitivity analysis of the modal data obtained from experimental modal analysis(EMA) and analytical modal analysis(AMA). The adapted sensitivity analysis methods of the algorithms are 1)eigensensitivity(EGNS) method. 2)frequency response function sensitivity(FRFS) method. 3)sensitivity based element-by-element method (SBEEM), Singular value decomposition(SVD) is used for performing eigenanalysis and parameter estimation in the updating process. Those algorithms are applied to finite element of a plate and the updating capability of each algorithm is compared in terms of accuracy. reliability and stability of the updating process. It is shown that the model updating method using frequency response function is superior to the other methods in view of various updating capabilities.

  • PDF

Fractionally Spaced Blind Equalization Using Singular Value Decomposition (특이값 분해를 이용한 블라인드 부분 간격 등화기)

  • Kim, Geumbee;Lee, Jeongwon;Nam, Haewoon;Park, Daeyoung
    • The Journal of Korean Institute of Communications and Information Sciences
    • /
    • v.41 no.9
    • /
    • pp.1041-1043
    • /
    • 2016
  • This letter proposes a new blind fractionally spaced equalization (FSE). The conventional linear program (LP) FSE reduces the degree of freedom (DOF) by abandoning many equalization filter taps, which causes severe performance degradations. We use singular value decomposition (SVD) to obtain the signal subspace and to fully utilize all samples for performance improvement. The proposed scheme has similar performance with the nuclear norm minimization and has as low complexity as the LP equalizer.

A Damping Distribution Method for Inverse Kinematics Problem Near Singular Configurations (특이점 근방에서 역 기구학 해를 구하기 위한 자동 감쇄 분배 방법)

  • 성영휘
    • Journal of Institute of Control, Robotics and Systems
    • /
    • v.4 no.6
    • /
    • pp.780-785
    • /
    • 1998
  • In this paper, it is shown that the conventional methods for dealing with the singularity problem of a manipulator can be generalized as a local minimization problem with differently weighted objective functions. A new damping method proposed in this article automatically determines the damping amounts for singular values, which are inversely proportional to the magnitude of the singular values. Furthermore, this can be done without explicitly computing the singular values. The proposed method can be applied to all the manipulators with revolute joints.

  • PDF