Search results for: non-negative matrix factorization

Authors: Hironori Karachi, Haruka Yamashita

Abstract:

Recently, the system of quick response (QR) code is getting popular. Many companies introduce new QR code payment services and the services are competing with each other to increase the number of users. For increasing the number of users, we should grasp the difference of feature of the demographic information, usage information, and value of users between services. In this study, we conduct an analysis of real-world data provided by Nomura Research Institute including the demographic data of users and information of users’ usages of two services; LINE Pay, and PayPay. For analyzing such data and interpret the feature of them, Nonnegative Matrix Factorization (NMF) is widely used; however, in case of the target data, there is a problem of the missing data. EM-algorithm NMF (EMNMF) to complete unknown values for understanding the feature of the given data presented by matrix shape. Moreover, for comparing the result of the NMF analysis of two matrices, there is Discriminant NMF (DNMF) shows the difference of users features between two matrices. In this study, we combine EMNMF and DNMF and also analyze the target data. As the interpretation, we show the difference of the features of users between LINE Pay and Paypay.

Keywords: data science, non-negative matrix factorization, missing data, quality of services

Procedia PDF Downloads 131

Authors: Kazuyoshi Mori

Abstract:

Examples of parameterization of stabilizing controllers that require only one of right-/left-coprime factorizations are presented. One parameterization method requires one side coprime factorization. The other requires no coprime factorization. The methods are based on the factorization approach so that a number of models can be applied the method we use in this paper.

Keywords: parametrization, coprime factorization, factorization approach, linear systems

Procedia PDF Downloads 373

Authors: Ana Serafimovic, Karthik Devarajan

Abstract:

Non-negative matrix factorization approximates a high dimensional non-negative matrix V as the product of two non-negative matrices, W and H, and allows only additive linear combinations of data, enabling it to learn parts with representations in reality. It has been successfully applied in the analysis and interpretation of high dimensional data arising in neuroscience, computational biology, and natural language processing, to name a few. The objective of this paper is to assess a hybrid algorithm for non-negative matrix factorization with multiplicative updates. The method aims to minimize the symmetric version of Kullback-Leibler divergence known as intrinsic information and assumes that the noise is signal-dependent and that it originates from an arbitrary distribution from the exponential family. It is a generalization of currently available algorithms for Gaussian, Poisson, gamma and inverse Gaussian noise. We demonstrate the potential usefulness of the new generalized algorithm by comparing its performance to the baseline methods which also aim to minimize symmetric divergence measures.

Keywords: non-negative matrix factorization, dimension reduction, clustering, intrinsic information, symmetric information divergence, signal-dependent noise, exponential family, generalized Kullback-Leibler divergence, dual divergence

Procedia PDF Downloads 246

Authors: Ric Joseph R. Murillo, Edna N. Gueco, Dennis I. Merino

Abstract:

Let F be C or R, where C and R are the set of complex numbers and real numbers, respectively, and n be a natural number. An n-by-n matrix A over the field F is called an α-involutory matrix or an α-involution if there exists an α in the field such that the square of the matrix is equal to αI, where I is the n-by-n identity matrix. If α is a complex number or a nonnegative real number, then an n-by-n matrix A over the field F can be written as a sum of n-by-n α-involutory matrices over the field F if and only if the trace of that matrix is an integral multiple of the square root of α. Meanwhile, if α is a negative real number, then a 2n-by-2n matrix A over R can be written as a sum of 2n-by-2n α-involutory matrices over R if and only the trace of the matrix is zero. Some other properties of α-involutory matrices are also determined

Keywords: α-involutory Matrices, sum of α-involutory Matrices, Trace, Matrix Theory

Procedia PDF Downloads 198

Authors: Brando Vagenende, Marie-Anne Guerry

Abstract:

For stochastic matrices of any order, the geometric description of the convex set of eigenvalues is completely known. The purpose of this study is to investigate the subset of the monotone matrices. This type of matrix appears in contexts such as intergenerational occupational mobility, equal-input modeling, and credit ratings-based systems. Monotone matrices are stochastic matrices in which each row stochastically dominates the previous row. The monotonicity property of a stochastic matrix can be expressed by a nonnegative lower-order matrix with the same eigenvalues as the original monotone matrix (except for the eigenvalue 1). Specifically, the aim of this research is to focus on the properties of eigenvalues of monotone matrices. For those matrices up to order 3, there already exists a complete description of the convex set of eigenvalues. For monotone matrices of order at least 4, this study gives, through simulations, more insight into the geometric description of their eigenvalues. Furthermore, this research treats in a geometric and algebraic way the properties of eigenvalues of monotone matrices of order at least 4.

Keywords: eigenvalues of matrices, finite Markov chains, monotone matrices, nonnegative matrices, stochastic matrices

Procedia PDF Downloads 80

Authors: Alfi Y. Zakiyyah, Hanni Garminia, M. Salman, A. N. Irawati

Abstract:

In the terminology of the hypergraph, there is a relation with the terminology graph. In the theory of graph, the edges connected two vertices. In otherwise, in hypergraph, the edges can connect more than two vertices. There is representation matrix of a graph such as adjacency matrix, Laplacian matrix, and incidence matrix. The adjacency matrix is symmetry matrix so that all eigenvalues is real. This matrix is a nonnegative matrix. The all diagonal entry from adjacency matrix is zero so that the trace is zero. Another representation matrix of the graph is the Laplacian matrix. Laplacian matrix is symmetry matrix and semidefinite positive so that all eigenvalues are real and non-negative. According to the spectral study in the graph, some that result is generalized to hypergraph. A hypergraph can be represented by a matrix such as adjacency, incidence, and Laplacian matrix. Throughout for this term, we use Laplacian matrix to represent a complete tripartite hypergraph. The aim from this research is to determine second smallest eigenvalues from this matrix and find a relation this eigenvalue with the connectivity of that hypergraph.

Keywords: connectivity, graph, hypergraph, Laplacian matrix

Procedia PDF Downloads 488

Authors: Adel Mohsen

Abstract:

A direct method based on the good Broyden's updates for evaluating the inverse of a nonsingular square matrix of full rank and solving related system of linear algebraic equations is studied. For a matrix A of order n whose LU-decomposition is A = LU, the multiplication count is O (n3). This includes the evaluation of the LU-decompositions of the inverse, the lower triangular decomposition of A as well as a “reduced matrix inverse”. If an explicit value of the inverse is not needed the order reduces to O (n3/2) to compute to compute inv(U) and the reduced inverse. For a symmetric matrix only O (n3/3) operations are required to compute inv(L) and the reduced inverse. An example is presented to demonstrate the capability of using the reduced matrix inverse in treating ill-conditioned systems. Besides the simplicity of Broyden's update, the method provides a mean to exploit the possible sparsity in the matrix and to derive a suitable preconditioner.

Keywords: Broyden's updates, matrix inverse, inverse factorization, solution of linear algebraic equations, ill-conditioned matrices, preconditioning

Procedia PDF Downloads 479

Authors: Benyelles Zakaria, Yousfi Djaafar, Karoui Moussa Sofiane

Abstract:

Agriculture is fundamental and remains an important objective in the Algerian economy, based on traditional techniques and structures, it generally has a purpose of consumption. Collection of agricultural statistics in Algeria is done using traditional methods, which consists of investigating the use of land through survey and field survey. These statistics suffer from problems such as poor data quality, the long delay between collection of their last final availability and high cost compared to their limited use. The objective of this work is to develop a processing chain for a reliable inventory of agricultural land by trying to develop and implement a new method of extracting information. Indeed, this methodology allowed us to combine data from remote sensing and field data to collect statistics on areas of different land. The contribution of remote sensing in the improvement of agricultural statistics, in terms of area, has been studied in the wilaya of Sidi Bel Abbes. It is in this context that we applied a method for extracting information from satellite images. This method is called the non-negative matrix factorization, which does not consider the pixel as a single entity, but will look for components the pixel itself. The results obtained by the application of the MNF were compared with field data and the results obtained by the method of maximum likelihood. We have seen a rapprochement between the most important results of the FMN and those of field data. We believe that this method of extracting information from satellite data leads to interesting results of different types of land uses.

Keywords: blind source separation, hyper-spectral image, non-negative matrix factorization, remote sensing

Procedia PDF Downloads 423

Authors: Yixun Shi

Abstract:

Many interior point methods for convex programming solve an (n+m)x(n+m)linear system in each iteration. Many implementations solve this system in each iteration by considering an equivalent mXm system (4) as listed in the paper, and thus the job is reduced into solving the system (4). However, the system(4) has to be solved exactly since otherwise the error would be entirely passed onto the last m equations of the original system. Often the Cholesky factorization is computed to obtain the exact solution of (4). One Cholesky factorization is to be done in every iteration, resulting in higher computational costs. In this paper, two iterative methods for solving linear systems using vector division are combined together and embedded into interior point methods. Instead of computing one Cholesky factorization in each iteration, it requires only one Cholesky factorization in the entire procedure, thus significantly reduces the amount of computation needed for solving the problem. Based on that, a hybrid algorithm for solving convex programming problems is proposed.

Keywords: convex programming, interior point method, linear systems, vector division

Procedia PDF Downloads 402

Authors: Moulana Mohammed

Abstract:

Topic models are widely used in building clusters of documents for more than a decade, yet problems occurring in choosing optimal number of topics. The main problem is the lack of a stable metric of the quality of topics obtained during the construction of topic models. The authors analyzed from previous works, most of the models used in determining the number of topics are non-parametric and quality of topics determined by using perplexity and coherence measures and concluded that they are not applicable in solving this problem. In this paper, we used the parametric method, which is an extension of the traditional topic model with visual access tendency for visualization of the number of topics (clusters) to complement clustering and to choose optimal number of topics based on results of cluster validity indices. Developed hybrid topic models are demonstrated with different Twitter datasets on various topics in obtaining the optimal number of topics and in measuring the quality of clusters. The experimental results showed that the Visual Non-negative Matrix Factorization (VNMF) topic model performs well in determining the optimal number of topics with interactive visualization and in performance measure of the quality of clusters with validity indices.

Keywords: interactive visualization, visual mon-negative matrix factorization model, optimal number of topics, cluster validity indices, Twitter data clustering

Procedia PDF Downloads 134

Authors: Caihong Sun, Xizi Zhang

Abstract:

Rating prediction is an important problem for recommender systems. The task is to predict the rating for an item that a user would give. Most of the existing algorithms for the task ignore the effect of negative ratings rated by users on items, but the negative ratings have a significant impact on users’ purchasing decisions in practice. In this paper, we present a rating prediction algorithm based on factorization machines that consider the effect of negative ratings inspired by Loss Aversion theory. The aim of this paper is to develop a concave and a convex negative disgust function to evaluate the negative ratings respectively. Experiments are conducted on MovieLens dataset. The experimental results demonstrate the effectiveness of the proposed methods by comparing with other four the state-of-the-art approaches. The negative ratings showed much importance in the accuracy of ratings predictions.

Keywords: factorization machines, feature engineering, negative ratings, recommendation systems

Procedia PDF Downloads 242

Authors: Martins Olatunbosun Babatunde, Muhammed Bashir Muazu, Emmanuel Adewale Adedokun

Abstract:

This paper presents the development of a double-loop trajectory-tracking controller for the ball and plate system (BPS) using the Normalized Right Coprime Factorization (NRCF) scheme.The Linear Algebraic (LA) method is used to design the inner loop required to stabilize the ball, while H-infinity NRCF method, that involved the lead-lag compensator design approach, is used to develop the outer loop that controls the plate. Simulation results show that the plate was stabilized at 0.2989 seconds and the ball was able to settle after 0.9646 seconds, with a trajectory tracking error of 0.0036. This shows that the controller has good adaptability and robustness.

Keywords: ball and plate system, normalized right coprime factorization, linear algebraic method, compensator, controller, tracking.

Procedia PDF Downloads 141

Authors: Bharat Singh Om Prakash Vyas

Abstract:

Now a day’s applications deal with High Dimensional Data have tremendously used in the popular areas. To tackle with such kind of data various approached has been developed by researchers in the last few decades. To tackle with such kind of data various approached has been developed by researchers in the last few decades. One of the problems with the NMF approaches, its randomized valued could not provide absolute optimization in limited iteration, but having local optimization. Due to this, we have proposed a new approach that considers the initial values of the decomposition to tackle the issues of computationally expensive. We have devised an algorithm for initializing the values of the decomposed matrix based on the PSO (Particle Swarm Optimization). Through the experimental result, we will show the proposed method converse very fast in comparison to other row rank approximation like simple NMF multiplicative, and ACLS techniques.

Keywords: ALS, NMF, high dimensional data, RMSE

Procedia PDF Downloads 342

Authors: Linda Smail, Zineb Azouz

Abstract:

Given a Bayesian network relative to a set I of discrete random variables, we are interested in computing the probability distribution P(S) where S is a subset of I. The general idea is to write the expression of P(S) in the form of a product of factors where each factor is easy to compute. More importantly, it will be very useful to give an interpretation of each of the factors in terms of conditional probabilities. This paper considers a semantic interpretation of the factors involved in computing marginal probabilities in Bayesian networks. Establishing such a semantic interpretations is indeed interesting and relevant in the case of large Bayesian networks.

Keywords: Bayesian networks, D-Separation, level two Bayesian networks, factorization of computation

Procedia PDF Downloads 529

Authors: Artion Kashuri, Rozana Liko

Abstract:

In this paper, the authors establish an integral identity using delta differentiable functions. By applying this identity, some new results via a general class of convex functions with respect to two nonnegative functions on a time scale are given. Also, for suitable choices of nonnegative functions, some special cases are deduced. Finally, in order to illustrate the efficiency of our main results, some applications to special means are obtained as well. We hope that current work using our idea and technique will attract the attention of researchers working in mathematical analysis, mathematical inequalities, numerical analysis, special functions, fractional calculus, quantum mechanics, quantum calculus, physics, probability and statistics, differential and difference equations, optimization theory, and other related fields in pure and applied sciences.

Keywords: convex functions, Hermite-Hadamard inequality, special means, time scale

Procedia PDF Downloads 150

Authors: Kazuyoshi Mori

Abstract:

In this paper, we consider the parametrization of Anantharam’s model within the framework of the factorization approach. In the parametrization, we investigate the number of required parameters of Anantharam’s model. We consider single-input single-output systems in this paper. By the investigation, we find three cases that are (1) there exist plants which require only one parameter and (2) two parameters, and (3) the number of parameters is at most three.

Keywords: linear systems, parametrization, coprime factorization, number of parameters

Procedia PDF Downloads 213

Authors: Soyon Kim, Edward Kim

Abstract:

In order to retrieve information from the massive stream of songs in the music industry, music search by title, lyrics, artist, mood, and genre has become more important. Despite the subjectivity and controversy over the definition of music genres across different nations and cultures, automatic genre classification systems that facilitate the process of music categorization have been developed. Manual genre selection by music producers is being provided as statistical data for designing automatic genre classification systems. In this paper, an automatic music genre classification system utilizing non-negative matrix factorization (NMF) is proposed. Short-term characteristics of the music signal can be captured based on the timbre features such as mel-frequency cepstral coefficient (MFCC), decorrelated filter bank (DFB), octave-based spectral contrast (OSC), and octave band sum (OBS). Long-term time-varying characteristics of the music signal can be summarized with (1) the statistical features such as mean, variance, minimum, and maximum of the timbre features and (2) the modulation spectrum features such as spectral flatness measure, spectral crest measure, spectral peak, spectral valley, and spectral contrast of the timbre features. Not only these conventional basic long-term feature vectors, but also NMF based feature vectors are proposed to be used together for genre classification. In the training stage, NMF basis vectors were extracted for each genre class. The NMF features were calculated in the log spectral magnitude domain (NMF-LSM) as well as in the basic feature vector domain (NMF-BFV). For NMF-LSM, an entire full band spectrum was used. However, for NMF-BFV, only low band spectrum was used since high frequency modulation spectrum of the basic feature vectors did not contain important information for genre classification. In the test stage, using the set of pre-trained NMF basis vectors, the genre classification system extracted the NMF weighting values of each genre as the NMF feature vectors. A support vector machine (SVM) was used as a classifier. The GTZAN multi-genre music database was used for training and testing. It is composed of 10 genres and 100 songs for each genre. To increase the reliability of the experiments, 10-fold cross validation was used. For a given input song, an extracted NMF-LSM feature vector was composed of 10 weighting values that corresponded to the classification probabilities for 10 genres. An NMF-BFV feature vector also had a dimensionality of 10. Combined with the basic long-term features such as statistical features and modulation spectrum features, the NMF features provided the increased accuracy with a slight increase in feature dimensionality. The conventional basic features by themselves yielded 84.0% accuracy, but the basic features with NMF-LSM and NMF-BFV provided 85.1% and 84.2% accuracy, respectively. The basic features required dimensionality of 460, but NMF-LSM and NMF-BFV required dimensionalities of 10 and 10, respectively. Combining the basic features, NMF-LSM and NMF-BFV together with the SVM with a radial basis function (RBF) kernel produced the significantly higher classification accuracy of 88.3% with a feature dimensionality of 480.

Keywords: mel-frequency cepstral coefficient (MFCC), music genre classification, non-negative matrix factorization (NMF), support vector machine (SVM)

Procedia PDF Downloads 303

Authors: Alica Miller

Abstract:

A semiflow is a triple consisting of a Hausdorff topological space $X$, a commutative topological monoid $T$ and a continuous monoid action of $T$ on $X$. The acting monoid $T$ is usually either the discrete monoid $\N_0$ of nonnegative integers (in which case the semiflow can be defined as a pair $(X,f)$ consisting of a phase space $X$ and a continuous function $f:X\to X$), or the monoid $\R_+$ of nonnegative real numbers (the so-called one-parameter monoid). However, it turns out that there are real-life situations where it is useful to consider the acting monoids that are a combination of discrete and continuous monoids. That, for example, happens, when we are observing certain dynamical system at discrete moments, but after some time realize that it would be beneficial to continue our observations in real time. The acting monoid in that case would be $T=\{0, t_0, 2t_0, \dots, (n-1)t_0\} \cup [nt_0,\infty)$ with the operation and topology induced from real numbers. This partly explains the motivation for the level of generality which is pursued in our research. We introduce the PSP monoids, which include all but ``pathological'' monoids, and most of our statements hold for them. The topic of our presentation are some recent results about chaos-related properties in semiflows, indecomposability and sensitivity of semiflows in the described general context.

Keywords: chaos, indecomposability, PSP monoids, semiflow, sensitivity

Procedia PDF Downloads 285

Authors: Ogundele Lasun Tunde, Ayeku Oluwagbemiga Patrick

Abstract:

Heavy metals are released into the sediments in aquatic environment from both natural and anthropogenic sources and they are considered as worldwide issue due to their deleterious ecological risks and food chain disruption. In this study, sediments samples were collected at three major sites (Awoye, Abereke and Ayetoro) along Ondo coastal area using VanVeen grab sampler. The concentrations of As, Cd, Cr, Cu, Fe, Mn, Ni, Pb, V and Zn were determined by employing Atomic Absorption Spectroscopy (AAS). The combined concentrations data were subjected to Positive Matrix Factorization (PMF) receptor approach for source identification and apportionment. The probable risks that might be posed by heavy metals in the sediment were estimated by potential and integrated ecological risks indices. Among the measured heavy metals, Fe had the average concentrations of 20.38 ± 2.86, 23.56 ± 4.16 and 25.32 ± 4.83 lg/g at Abereke, Awoye and Ayetoro sites, respectively. The PMF resulted in identification of four sources of heavy metals in the sediments. The resolved sources and their percentage contributions were oil exploration (39%), industrial waste/sludge (35%), detrital process (18%) and Mn-sources (8%). Oil exploration activities and industrial wastes are the major sources that contribute heavy metals into the coastal sediments. The major pollutants that posed ecological risks to the local aquatic ecosystem are As, Pb, Cr and Cd (40 B Ei ≤ 80) classifying the sites as moderate risk. The integrate risks values of Awoye, Abereke and Ayetoro are 231.2, 234.0 and 236.4, respectively suggesting that the study areas had a moderate ecological risk. The study showed the suitability of PMF receptor model for source identification of heavy metals in the sediments. Also, the intensive anthropogenic activities and natural sources could largely discharge heavy metals into the study area, which may increase the heavy metal contents of the sediments and further contribute to the associated ecological risk, thus affecting the local aquatic ecosystem.

Keywords: positive matrix factorization, sediments, heavy metals, sources, ecological risks

Procedia PDF Downloads 21

Authors: M. Hadji, N. Chiker, Y. Hadji, A. Haddad

Abstract:

In this paper, we report for the first time on the synthesis and characterization of novel MAX phases (Ti3SiC2, Ti2AlC) reinforced Ni-matrix and Ti2AlC reinforced Al-matrix. The stability of MAX phases in Al-matrix and Ni-matrix at a temperature of 985°C has been investigated. All the composites were cold pressed and sintered at a temperature of 985°C for 20min in H2 environment, except (Ni/Ti3SiC2) who was sintered at 1100°C for 1h.Microstructure analysis by scanning electron microscopy and phase analysis by X-Ray diffraction confirmed that there was minimal interfacial reaction between MAX particles and Ni, thus Al/MAX samples shown that MAX phases was totally decomposed at 985°C.The Addition of MAX enhanced the Al-matrix and Ni-matrix.

Keywords: MAX phase, microstructures, composites, hardness, SEM

Procedia PDF Downloads 347

Authors: Renata Masarova, Bohuslava Juhasova, Martin Juhas, Zuzana Sutova

Abstract:

In dynamic system theory a mathematical model is often used to describe their properties. In order to find a transfer matrix of a dynamic system we need to calculate an inverse matrix. The paper contains the fusion of the classical theory and the procedures used in the theory of automated control for calculating the inverse matrix. The final part of the paper models the given problem by the Matlab.

Keywords: dynamic system, transfer matrix, inverse matrix, modeling

Procedia PDF Downloads 515

Authors: Kirill Trapezon, Alexandr Trapezon

Abstract:

A problem is formulated for the natural oscillations of a circular plate of linearly variable thickness on the basis of the symmetry method. The equations of natural frequencies and forms for a plate are obtained, providing that it is rigidly fixed along the inner contour. The first three eigenfrequencies are calculated, and the eigenmodes of the oscillations of the acoustic element are constructed. An algorithm for applying the symmetry method and the factorization method for solving problems in the theory of oscillations for plates of variable thickness is shown. The effectiveness of the approach is demonstrated on the basis of comparison of known results and those obtained in the article. It is shown that the results are more accurate and reliable.

Keywords: vibrations, plate, method of symmetries, differential equation, factorization, approximation

Procedia PDF Downloads 262

Authors: Danny Manuel Calvo Velasco, Robert Sanchez Cano

Abstract:

By using the transference matrix formalism, in this work, it is presented the study of the optical properties of the 1D graded structure, constructed by multiple bi-layers of dielectric and air, considering a redistribution of the material lengths following an arithmetic progression as a function of two parameters. It is presented a factorization for the transference matrices for the graded structure, which allows the interpretation of their optical properties in terms of the properties of simpler structures. It is shown that the graded structure presents new transmission peaks, which can be controlled by the parameter values located in frequencies for which a periodic system has a photonic bandgap. This result is extended to the case of a photonic crystal for which the unitary cell is the proposed graded structure, showing new transmission bands which are due to the multiple new sub-structures present in the system. Also, for the TE polarization, it is observed transmission bands' low frequencies which present low variation of its width and position with the incidence angle. It is expected that these results could guide a route in the design of new photonic devices.

Keywords: graded, material redistribution, photonic system, transference matrix

Procedia PDF Downloads 139

Authors: Boris Melnikov, Dmitrii Chaikovskii, Elena Melnikova

Abstract:

The traveling salesman problem (TSP) is a well-known optimization problem that seeks to find the shortest possible route that visits a set of points and returns to the starting point. In this paper, we apply some heuristics of the TSP for the task of restoring the DNA matrix. This restoration problem is often considered in biocybernetics. For it, we must recover the matrix of distances between DNA sequences if not all the elements of the matrix under consideration are known at the input. We consider the possibility of using this method in the testing of distance calculation algorithms between a pair of DNAs to restore the partially filled matrix.

Keywords: optimization problems, DNA matrix, partially filled matrix, traveling salesman problem, heuristic algorithms

Procedia PDF Downloads 150

Authors: Kourosh Modarresi

Abstract:

The Netflix problem has brought the topic of “Recommendation Systems” into the mainstream of computer science, mathematics, and statistics. Though much progress has been made, the available algorithms do not obtain satisfactory results. The success of these algorithms is rarely above 5%. This work is based on the belief that the main challenge is to come up with “scalable personalization” models. This paper uses an adaptive regularization of inverse singular value decomposition (SVD) that applies adaptive penalization on the singular vectors. The results show far better matching for recommender systems when compared to the ones from the state of the art models in the industry.

Keywords: convex optimization, LASSO, regression, recommender systems, singular value decomposition, low rank approximation

Procedia PDF Downloads 455

Authors: Shih Yu Pan, Pao Chen Hung, Chuan Yao Lin, Charles C.-K. Chou, Yu Chi Lin, Kai Hsien Chi

Abstract:

This study analyzed 16 atmospheric PAHs species which were controlled by USEPA and IARC. To measure the concentration of PAHs, four rural sampling sites and two urban sampling sites were selected in Central Taiwan during spring and summer. In central Taiwan, the rural sampling stations were located in the downstream of Da-An River, Da-Jang River, Wu River and Chuo-shui River. On the other hand, the urban sampling sites were located in Taichung district and close to the roadside. Ambient air samples of both vapor phase and particle phase of PAHs compounds were collected using high volume sampling trains (Analitica). The sampling media were polyurethane foam (PUF) with XAD2 and quartz fiber filters. Diagnostic ratio, Principal component analysis (PCA), Positive Matrix Factorization (PMF) models were used to evaluate the apportionment of PAHs in the atmosphere and speculate the relative contribution of various emission sources. Because of the high temperature and low wind speed, high PAHs concentration in the atmosphere was observed. The total PAHs concentration, especially in vapor phase, had significant change during summer. During the sampling periods the total PAHs concentration of atmospheric at four rural and two urban sampling sites in spring and summer were 3.70±0.40 ng/m3,3.40±0.63 ng/m3,5.22±1.24 ng/m3,7.23±0.37 ng/m3,7.46±2.36 ng/m3,6.21±0.55 ng/m3　; 15.0± 0.14 ng/m3,18.8±8.05 ng/m3,20.2±8.58 ng/m3,16.1±3.75 ng/m3,29.8±10.4 ng/m3,35.3±11.8 ng/m3, respectively. In order to identify PAHs sources, we used diagnostic ratio to classify the emission sources. The potential sources were diesel combustion and gasoline combustion in spring and summer, respectively. According to the principal component analysis (PCA), the PC1 and PC2 had 23.8%, 20.4% variance and 21.3%, 17.1% variance in spring and summer, respectively. Especially high molecular weight PAHs (BaP, IND, BghiP, Flu, Phe, Flt, Pyr) were dominated in spring when low molecular weight PAHs (AcPy, Ant, Acp, Flu) because of the dominating high temperatures were dominated in the summer. Analysis by using PMF model found the sources of PAHs in spring were stationary sources (34%), vehicle emissions (24%), coal combustion (23%) and petrochemical fuel gas (19%), while in summer the emission sources were petrochemical fuel gas (34%), the natural environment of volatile organic compounds (29%), coal combustion (19%) and stationary sources (18%).

Keywords: PAHs, source identification, diagnostic ratio, principal component analysis, positive matrix factorization

Procedia PDF Downloads 267

Authors: Jiao Sun, Li Pan, Shijun Liu

Abstract:

Recommender systems have been widely used in contemporary industry, and plenty of work has been done in this field to help users to identify items of interest. Collaborative Filtering (CF, for short) algorithm is an important technology in recommender systems. However, less work has been done in automobile recommendation system with the sharp increase of the amount of automobiles. What’s more, the computational speed is a major weakness for collaborative filtering technology. Therefore, using MapReduce framework to optimize the CF algorithm is a vital solution to this performance problem. In this paper, we present a recommendation of the users’ comment on industrial automobiles with various properties based on real world industrial datasets of user-automobile comment data collection, and provide recommendation for automobile providers and help them predict users’ comment on automobiles with new-coming property. Firstly, we solve the sparseness of matrix using previous construction of score matrix. Secondly, we solve the data normalization problem by removing dimensional effects from the raw data of automobiles, where different dimensions of automobile properties bring great error to the calculation of CF. Finally, we use the MapReduce framework to optimize the CF algorithm, and the computational speed has been improved times. UV decomposition used in this paper is an often used matrix factorization technology in CF algorithm, without calculating the interpolation weight of neighbors, which will be more convenient in industry.

Keywords: collaborative filtering, recommendation, data normalization, mapreduce

Procedia PDF Downloads 217

Authors: Malika Yaici, Kamel Hariche

Abstract:

In control engineering, systems described by matrix fractions are studied through properties of block roots, also called solvents. These solvents are usually dealt with in a block Vandermonde matrix form. Inverses and determinants of Vandermonde matrices and block Vandermonde matrices are used in solving problems of numerical analysis in many domains but require costly computations. Even though Vandermonde matrices are well known and method to compute inverse and determinants are many and, generally, based on interpolation techniques, methods to compute the inverse and determinant of a block Vandermonde matrix have not been well studied. In this paper, some properties of these matrices and iterative algorithms to compute the determinant and the inverse of a block Vandermonde matrix are given. These methods are deducted from the partitioned matrix inversion and determinant computing methods. Due to their great size, parallelization may be a solution to reduce the computations cost, so a parallelization of these algorithms is proposed and validated by a comparison using algorithmic complexity.

Keywords: block vandermonde matrix, solvents, matrix polynomial, matrix inverse, matrix determinant, parallelization

Procedia PDF Downloads 239

Authors: Kazuyoshi Mori

Abstract:

In this paper, we consider the parametrization of the discrete-time systems without the unit-delay element within the framework of the factorization approach. In the parametrization, we investigate the number of required parameters. We consider single-input single-output systems in this paper. By the investigation, we find, on the discrete-time systems without the unit-delay element, three cases that are (1) there exist plants which require only one parameter and (2) two parameters, and (3) the number of parameters is at most three.

Keywords: factorization approach, discrete-time system, parameterization of stabilizing controllers, system without unit-delay

Procedia PDF Downloads 240

Authors: Bipin Popat Sonawane

Abstract:

After 1988 Electron muon Collaboration (EMC) announcement of measurement of spin dependent structure function, it has been found that it has become a need to understand spin structure of a hadron. In the study of three-dimensional spin structure of a proton, we need to understand the foundation of quantum field theory in terms of electro-weak and strong theories using rigorous mathematical theories and models. In the process of understanding the inner dynamical stricture of proton we need understand the mathematical formalism in perturbative quantum chromodynamics (pQCD). In QCD processes like proton-proton collision at high energy we calculate cross section using conventional collinear factorization schemes. In this calculations, parton distribution functions (PDFs) and fragmentation function are used which provide the information about probability density of finding quarks and gluons ( partons) inside the proton and probability density of finding final hadronic state from initial partons. In transverse momentum dependent (TMD) PDFs and FFs, collectively called as TMDs, take an account for intrinsic transverse motion of partons. The TMD factorization in the calculation of cross sections provide a scheme of hadronic and partonic states in the given QCD process. In this study we review Transverse Momentum Dependent (TMD) factorization scheme using Collins-Soper-Sterman (CSS) Formalism. CSS formalism considers the transverse momentum dependence of the partons, in this formalism the cross section is written as a Fourier transform over a transverse position variable which has physical interpretation as impact parameter. Along with this we compare this formalism with improved CSS formalism. In this work we study the TMD evolution schemes and their comparison with other schemes. This would provide description in the process of measurement of transverse single spin asymmetry (TSSA) in hadro-production and electro-production of J/psi meson at RHIC, LHC, ILC energy scales. This would surely help us to understand J/psi production mechanism which is an appropriate test of QCD.

Keywords: QCD, PDF, TMD, CSS

Procedia PDF Downloads 69