• Title/Summary/Keyword: Matrix factorization

Search Result 309, Processing Time 0.021 seconds

Audio signal clustering and separation using a stacked autoencoder (복층 자기부호화기를 이용한 음향 신호 군집화 및 분리)

  • Jang, Gil-Jin
    • The Journal of the Acoustical Society of Korea
    • /
    • v.35 no.4
    • /
    • pp.303-309
    • /
    • 2016
  • This paper proposes a novel approach to the problem of audio signal clustering using a stacked autoencoder. The proposed stacked autoencoder learns an efficient representation for the input signal, enables clustering constituent signals with similar characteristics, and therefore the original sources can be separated based on the clustering results. STFT (Short-Time Fourier Transform) is performed to extract time-frequency spectrum, and rectangular windows at all the possible locations are used as input values to the autoencoder. The outputs at the middle, encoding layer, are used to cluster the rectangular windows and the original sources are separated by the Wiener filters derived from the clustering results. Source separation experiments were carried out in comparison to the conventional NMF (Non-negative Matrix Factorization), and the estimated sources by the proposed method well represent the characteristics of the orignal sources as shown in the time-frequency representation.

Generic Document Summarization using Coherence of Sentence Cluster and Semantic Feature (문장군집의 응집도와 의미특징을 이용한 포괄적 문서요약)

  • Park, Sun;Lee, Yeonwoo;Shim, Chun Sik;Lee, Seong Ro
    • Journal of the Korea Institute of Information and Communication Engineering
    • /
    • v.16 no.12
    • /
    • pp.2607-2613
    • /
    • 2012
  • The results of inherent knowledge based generic summarization are influenced by the composition of sentence in document set. In order to resolve the problem, this papser propses a new generic document summarization which uses clustering of semantic feature of document and coherence of document cluster. The proposed method clusters sentences using semantic feature deriving from NMF(non-negative matrix factorization), which it can classify document topic group because inherent structure of document are well represented by the sentence cluster. In addition, the method can improve the quality of summarization because the importance sentences are extracted by using coherence of sentence cluster and the cluster refinement by re-cluster. The experimental results demonstrate appling the proposed method to generic summarization achieves better performance than generic document summarization methods.

Vehicle Recognition using NMF in Urban Scene (도심 영상에서의 비음수행렬분해를 이용한 차량 인식)

  • Ban, Jae-Min;Lee, Byeong-Rae;Kang, Hyun-Chul
    • The Journal of Korean Institute of Communications and Information Sciences
    • /
    • v.37 no.7C
    • /
    • pp.554-564
    • /
    • 2012
  • The vehicle recognition consists of two steps; the vehicle region detection step and the vehicle identification step based on the feature extracted from the detected region. Features using linear transformations have the effect of dimension reduction as well as represent statistical characteristics, and show the robustness in translation and rotation of objects. Among the linear transformations, the NMF(Non-negative Matrix Factorization) is one of part-based representation. Therefore, we can extract NMF features with sparsity and improve the vehicle recognition rate by the representation of local features of a car as a basis vector. In this paper, we propose a feature extraction using NMF suitable for the vehicle recognition, and verify the recognition rate with it. Also, we compared the vehicle recognition rate for the occluded area using the SNMF(sparse NMF) which has basis vectors with constraint and LVQ2 neural network. We showed that the feature through the proposed NMF is robust in the urban scene where occlusions are frequently occur.

Positive and Negative Covariation Mechanism of Multiple Muscle Activities During Human Walking (보행 과정에서 발생하는 복합 근육 활성의 양성 및 음성 공변 메커니즘)

  • Kim, Yushin;Hong, Youngki
    • The Journal of the Korea Contents Association
    • /
    • v.18 no.1
    • /
    • pp.173-184
    • /
    • 2018
  • In human walking, muscle co-contraction which produces simultaneous activities of multiple muscles is important in motor control mechanism of the central nervous system. This study aims to understand positive and negative covariation mechanism of inter-muscle activities during walking. In this study, we measured electromyography (EMG) in leg muscles. To identify motor modules, we recored EMG from 4 leg muscles bilaterally (the tibialis anterior, medial gastrocnemius, rectus femoris and medial hamstring muscles) and performed non-negative matrix factorization (NMF) and principa component analysis (PCA). Then, we computed covariation values from various combinations between muscles or motor modules and used two-way repeated measures analysis of variance to identify significantly different covariation patterns between muscle combinations. As the results, we found significant differences between covariation values of muscle combinations (p < 0.05). muscle groups within the same motor modules produced the positive covariations. However, there were strong negative covariation between motor modules. There was negative covariation in all muscle combination. Stable inter-module negative covariation suggests that motor modules may be the control unit in the complex motor coordination.

Local Region Spectral Analysis for Performance Enhancement of Dementia Classification (인지증 판별 성능 향상을 위한 스펙트럼 국부 영역 분석 방법)

  • Park, Jun-Qyu;Baek, Seong-Joon
    • Journal of the Korea Academia-Industrial cooperation Society
    • /
    • v.12 no.11
    • /
    • pp.5150-5155
    • /
    • 2011
  • Alzheimer's disease (AD) and vascular dementia (VD) are the most common dementia. In this paper, we proposed a region selection for classification of AD, VD and normal (NOR) based on micro-Raman spectra from platelet. The preprocessing step is a smoothing followed by background elimination to the original spectra. Then we applied the minmax method for normalization. After the inspection of the preprocessed spectra, we found that 725-777, 1504-1592 and 1632-1700 cm1 regions are the most discriminative features in AD, VD and NOR spectra. We applied the feature transformation using PCA (principal component analysis) and NMF (nonnegative matrix factorization). The classification result of MAP(maximum a posteriori probability) involving 327 spectra transformed features using proposed local region showed about 92.8 % true classification average rate.

Comparison of independent component analysis algorithms for low-frequency interference of passive line array sonars (수동 선배열 소나의 저주파 간섭 신호에 대한 독립성분분석 알고리즘 비교)

  • Kim, Juho;Ashraf, Hina;Lee, Chong-Hyun;Cheong, Myoung Jun
    • The Journal of the Acoustical Society of Korea
    • /
    • v.38 no.2
    • /
    • pp.177-183
    • /
    • 2019
  • In this paper, we proposed an application method of ICA (Independent Component Analysis) to passive line array sonar to separate interferences from target signals in low frequency band and compared performance of three conventional ICA algorithms. Since the low frequency signals are received through larger bearing angles than other frequency bands, neighboring beam signals can be used to perform ICA as measurement signals of the ICA. We use three ICA algorithms such as Fast ICA, NNMF (Non-negative Matrix Factorization) and JADE (Joint Approximation Diagonalization of Eigen-matrices). Through experiments on real data obtained from passive line array sonar, it is verified that the interference can be separable from target signals by the suggested method and the JADE algorithm shows the best separation performance among the three algorithms.

Missing Data Modeling based on Matrix Factorization of Implicit Feedback Dataset (암시적 피드백 데이터의 행렬 분해 기반 누락 데이터 모델링)

  • Ji, JiaQi;Chung, Yeongjee
    • Journal of the Korea Institute of Information and Communication Engineering
    • /
    • v.23 no.5
    • /
    • pp.495-507
    • /
    • 2019
  • Data sparsity is one of the main challenges for the recommender system. The recommender system contains massive data in which only a small part is the observed data and the others are missing data. Most studies assume that missing data is randomly missing from the dataset. Therefore, they only use observed data to train recommendation model, then recommend items to users. In actual case, however, missing data do not lost randomly. In our research, treat these missing data as negative examples of users' interest. Three sample methods are seamlessly integrated into SVD++ algorithm and then propose SVD++_W, SVD++_R and SVD++_KNN algorithm. Experimental results show that proposed sample methods effectively improve the precision in Top-N recommendation over the baseline algorithms. Among the three improved algorithms, SVD++_KNN has the best performance, which shows that the KNN sample method is a more effective way to extract the negative examples of the users' interest.

A Comparative Analysis of Personalized Recommended Model Performance Using Online Shopping Mall Data (온라인 쇼핑몰 데이터를 이용한 개인화 추천 모델 성능 비교 분석)

  • Oh, Jaedong;Oh, Ha-young
    • Journal of the Korea Institute of Information and Communication Engineering
    • /
    • v.26 no.9
    • /
    • pp.1293-1304
    • /
    • 2022
  • The personalization recommendation system means analyzing each individual's interests or preferences and recommending information or products accordingly. These personalized recommendations can reduce the time consumers spend searching for information by accessing the products they need more quickly, and companies can increase corporate profits by recommending appropriate products that meet their needs. In this study, products are recommended to consumers using collaborative filtering, matrix factorization, and deep learning, which are representative personalization recommendation techniques. To this end, the data set after purchasing shopping mall products, which is raw data, is pre-processed in the form of transmitting the data set to the input of the recommended system, and the pre-processed data set is analyzed from various angles. In addition, each model performs verification and performance comparison on the recommended results, and explores the model with optimal performance, suggesting which model should be used when building the recommendation system at the mall.

Low-complexity Sensor Selection Based on QR factorization (QR 분해에 기반한 저 복잡도 센서 선택 알고리즘)

  • Yoon Hak, Kim
    • Journal of the Korea Institute of Information and Communication Engineering
    • /
    • v.27 no.1
    • /
    • pp.103-108
    • /
    • 2023
  • We study the problem of selecting a subset of sensor nodes in sensor networks in order to maximize the performance of parameter estimation. To achieve a low-complexity sensor selection algorithm, we propose a greedy iterative algorithm that allows us to select one sensor node at a time so as to maximize the log-determinant of the inverse of the estimation error covariance matrix without resort to direct minimization of the estimation error. We apply QR factorization to the observation matrix in the log-determinant to derive an analytic selection rule which enables a fast selection of the next node at each iteration. We conduct the extensive experiments to show that the proposed algorithm offers a competitive performance in terms of estimation performance and complexity as compared with previous sensor selection techniques and provides a practical solution to the selection problem for various network applications.

MULTI-BLOCK BOUNDARY VALUE METHODS FOR ORDINARY DIFFERENTIAL AND DIFFERENTIAL ALGEBRAIC EQUATIONS

  • OGUNFEYITIMI, S.E.;IKHILE, M.N.O.
    • Journal of the Korean Society for Industrial and Applied Mathematics
    • /
    • v.24 no.3
    • /
    • pp.243-291
    • /
    • 2020
  • In this paper, multi-block generalized backward differentiation methods for numerical solutions of ordinary differential and differential algebraic equations are introduced. This class of linear multi-block methods is implemented as multi-block boundary value methods (MB2 VMs). The root distribution of the stability polynomial of the new class of methods are determined using the Wiener-Hopf factorization of a matrix polynomial for the purpose of their correct implementation. Numerical tests, showing the potential of such methods for output of multi-block of solutions of the ordinary differential equations in the new approach are also reported herein. The methods which output multi-block of solutions of the ordinary differential equations on application, are unlike the conventional linear multistep methods which output a solution at a point or the conventional boundary value methods and multi-block methods which output only a block of solutions per step. The MB2 VMs introduced herein is a novel approach at developing very large scale integration methods (VLSIM) in the numerical solution of differential equations.