Separation Matrix Research Articles

Face recognition using non-negative matrix factorization with a single sample per person in a large database

There are various face recognition techniques in literature, which are faced with challenges such as occlusion, pose variation, illumination, and facial expressions. Existing methods often perform well when their database is small, or multiple samples per person exist. However, face recognition methods with just one reference sample per person may not work well, especially on a large database. To address this problem, this paper proposes a scheme to extract features from facial images. Using Non-negative Matrix Factorization (NMF), basic features are extracted from the face structure. The matrix of images is decomposed into basis matrix (W) and weight matrix (H). The basis matrix contains several versions of mouths, noses and other facial parts, where the various versions are in different locations or forms. Hence, to recognize a facial image in the database, searching is done on the weight matrices feature set. In this research, to more precisely form the structural elements, a separate basis matrix is constructed for the upper and lower parts of the facial images from the database. Also the images are enhanced using pre-processing techniques including histogram equalization, image intensity, and contrast limited adaptive histogram. The FERET database with 990 single images per person was used to evaluate the proposed method. Experimental results show that the proposed method can achieve a recognition rate close to 93%.

Motivos contenidos en narraciones tradicionales orales de los kari'ña, pemón y ye'kwana

Este estudio está circunscrito al contexto de tres etnias de la familia caribe, sub-rama norte, específicamente kari’ña, pemón y ye’kwana, todas radicadas dentro de los confines territoriales de Venezuela; sobre estas en conjunto ---en un marco etnolingüístico--- se ha realizado la indización de los elementos léxicos identificatorios de los ‘motivos’ (Motif-Index of Folk-Literature , Stith Thompson, 1950-1955) contenidos en narraciones orales utilizadas por estos pueblos originarios para tramitir sus tradiciones de una a otra generación. Thompson (1946, pág. 415; 1950-1955, pág. 1137) define ‘motivo’ como “el elemento más pequeño de una narración que tiene el poder de persistir en la tradición”. El índice de motivos, o IM, resultante deriva de 8 narraciones, etiquetadas archivos-texto 1 a 8: 2 de ellas correspondientes a los kari’ña, 2 a los pemón, y 4 a los ye’kwana. En los 8 archivos-texto se indizaron 31 motivos con referencia a entes humanos y no humanos, con rasgos físicos y comportamientos fantásticos, mayoritariamente imaginarios. También, se incluyeron motivos que aluden actividades realizadas por tales seres, para las cuales se requieren aptitudes e instrumentos de los que no disponen las gentes comunes. Los motivos indizados fueron presentados tanto contextualizados como en matrices separadas cuyos descriptores encajan dentro de tipos y subtipos definidos por Thompson(op.cit.); Valory (1967); y Wilbert (1979,1975). Las narraciones utilizadas fueron recopiladas por el autor en kari’ña, pemón y ye’kwana, y trascritas por este como parte de sus investigaciones sobre la morfosintaxis de estas lenguas. Sin embargo, para este estudio en particular, el autor se apoyó en segmantos de las traducciones al español de dichas narraciones previamente publicadas por Civrieux (1970, 1976); Armellada y Bentivenga (1980); Guevara (2011); Armellada (1973) y Arreaza Adam (2005). El IM constituido ha proporcionado datos para una adicional primera aproximación a la caracterización etnocultural de estos pueblos caribes del norte.

Efficient Clustering of Non-coherent and Coherent Components Regardless of Sources’ Powers for 2D DOA Estimation

Conventional decorrelation techniques that resolve all signals simultaneously are not efficient in mixture scenarios of non-coherent and coherent signals. In newer methods for one-dimensional arrays, non-coherent signals and coherent groups are resolved separately. However, employing an unreliable and non-adaptive threshold is the most significant disadvantage of these methods. On the other hand, they cannot be implemented for two-dimensional arrays. To deal with these issues, the signals separation using k-medoids clustering (SSKMC) algorithm was presented. Although the SSKMC algorithm does not have any of the shortcomings mentioned above, it relies on a basic limiting assumption that the sources should be equi-power. Therefore, the practical application of the SSKMC algorithm is facing a serious problem. In this paper, the SSKMC algorithm is extended so that it can be used even if the sources’ powers are not the same. First, the two-dimensional array is divided into several parallel linear sub-arrays. Then, by defining a components separation matrix, and employing its eigenvalues, the non-coherent and coherent components are identified. The effectiveness of the proposed solution is proven by mathematical facts. Simulation results verify the proofs and the benefit of the proposed solution.

Hybrid monolith assisted magnetic ion-imprinted polymer extraction coupled with ICP-MS for determination of trace Au(III) in environmental and mineral samples

In this work, by introducing solvothermal method for generating Fe3O4 nanoparticles and “one-pot” method for synthesizing organic–inorganic hybrid monolithic vinyl functionalized Fe3O4 (Fe3O4@HM), a facile procedure was proposed for preparing poly(1-vinylimidazole) (PVIM) functionalized gold ion imprinted polymer based MNPs@IIP (MNPs@HM@PVIM-Au-IIP). The as-prepared magnetic nanocomposites were characterized by scanning electron microscopy (SEM), transmission electron microscope (TEM), Fourier transform infrared spectroscopy (FT-IR), X-ray diffraction (XRD), thermogravimetric analysis (TGA), and vibrating sample magnetometry (VSM), etc. After the optimum conditions, including solution pH, adsorption time and sample volume, eluent composition and volume, and elution time were investigated, this MNPs@IIP was employed as ion imprinted polymer magnetic solid phase extraction (II-MSPE) matrix for separation and enrichment of trace gold ions followed by inductively coupled plasma mass spectrometric (ICP-MS) determination with excellent selectivity and sensitivity, and strong anti-interference ability. These merits were benefited from the formation of hybrid monolithic interlayer enduing the resultant MNPs@HM@PVIM-Au-IIP with good chemical and mechanical stability, as well as the dimensional homogeneity due to the solvothermal-generated Fe3O4 nanoparticles as the cores. The applicability superiority of this MNPs@IIP was confirmed by the rapid and efficient extraction of trace level Au(III) in environment water, sediment, soil, coal and ore samples with satisfactory results. The reliability of this II-MSPE-ICP-MS method was also verified by using corresponding Certified Standard Reference materials and spiking recovery of the real samples analyzed.

A Bayesian Factor Model for Spatial Panel Data with a Separable Covariance Approach

A hierarchical Bayesian factor model for multivariate spatially and temporally correlated data is proposed. This method searches factor scores incorporating a dependence within observations due to both a geographical and a temporal structure and it is an extension of a model proposed by Mezzetti (2012) using the results of a separable covariance matrix for the spatial panel data as in Leorato and Mezzetti (2016). A Gibbs sampling algorithm is implemented to sample from the posterior distributions. We illustrate the benefit and the performance of our model by analyzing death rates for different diseases together with some socio-economical and behavioural indicators and by analyzing simulated data.

Identification of sustainability indicators in the civil aviation sector in Dubai: a stakeholders’ perspective

PurposeThe civil aviation sector in Dubai has been contributing to its economic development by enhancing transportation services, allowing for the establishment of trade links, promoting tourism and providing job opportunities. Despite being a thriving sector of economic activity (Dubai was the third busiest airport in the world in 2018), a common understanding among the stakeholders on the sector’s long-term sustainability strategy is missing. The purpose of this paper is to contribute to the identification of important sustainability indicators in the civil aviation sector in Dubai.Design/methodology/approachThis study uses the stakeholder theory to identify sustainability indicators. It deploys a qualitative research design that collects data through multiple focus groups, and it accounts for the stakeholders’ views toward key sustainability indicators. This research classified stakeholders completely in a separate matrix preventing their integration and therefore the confounding of their views. The study addresses the complexity of sustainability in the sector and identifies key indicators based on comprehensive and valid data from different stakeholders. The study focuses separately on the three pillars of sustainability: social, economic and environmental sustainability and identifies sustainability indicators for each pillar from the views of stakeholders. The study uses thematic analysis to produce results by identifying themes according to the collected data.FindingsThe study finds that there is a lack of sustainability knowledge and awareness in the civil aviation sector in Dubai despite its buoyant growth and expansion. The results show that stakeholders highlight the importance of social sustainability indicators that value equitable development in the pursuit of business goals. They also value economic sustainability indicators that focus on changing consumption patterns and increasing preservation. They finally focus on environmental sustainability indicators that promote eco-efficiency and renewable energy. However, considerable differences exist among stakeholders regarding the importance and priority of their suggested indicators.Practical implicationsThe implications of the study are theoretical and practical. In the case of the former, it would contribute toward the reduction of uncertainty in the civil aviation sector in Dubai regarding the policy actions required to improve the sector’s sustainability. It is done by identifying the stakeholders’ views on the key sustainability indicators for each pillar of sustainability. In the case of practical implications, the outcomes of the study highlight specific directions that policy actions should take. Although the results are local, they can have tentative global policy implications. Finally, the results of the study are instrumental in addressing the sustainability debate within the civil aviation sector not only in Dubai but also globally.Originality/valueThe study extends existing research efforts to identify sustainability indicators in the civil aviation sector by considering the three different pillars of sustainability and by accounting for the diverse stakeholders’ views on the identification of key sustainability indicators. The study should be extended to account for the role of technological readiness in the region of other innovative approaches in the current civil aviation sectors.

Characterisation by Excitation-Emission Matrix Fluorescence Spectroscopy of Pigments in Mucus Secreted of Earthworm Eisenia foetida Exposed to Lead.

The earthworm exposed to toxics shows physiological responses as: avoidance and mucus secretion. Heavy metals are particularly toxic to earthworms and the mucus secretion has been considered as a defence mechanism against undesirable substance. The chromophores present in the mucus secretion of Eisenia foetida have been poorly studied. Mucus secretion of E. foetida was induced by PbCl2. High PbCl2 concentrations provoked abundant mucus secretion which showed fluorescence when illuminated by UV light. Dialysis membrane separation, UV Visible and Excitation-Emission Matrix Fluorescence (EEM) spectroscopy were used to characterise the fluorescent pigments. EEM spectroscopy analysis of the mucus secretion signalled three excitation-emission peaks at: 310/380nm, 370/520nm and 440/520nm. Two fluorophores were separated by dialysis. One of them matched the fluorescent compound riboflavin excitation-emission profile; the other is a protein with a peak 290/350nm. Native-PAGE electrophoresis was conducted to assess the riboflavin-biding ability of the coelomic fluid protein produced by Eisenia foetida showing a high riboflavin-biding ability.

Adaptive Speech Separation Based on Beamforming and Frequency Domain-Independent Component Analysis

Voice signals acquired by a microphone array often include considerable noise and mutual interference, seriously degrading the accuracy and speed of speech separation. Traditional beamforming is simple to implement, but its source interference suppression is not adequate. In contrast, independent component analysis (ICA) can improve separation, but imposes an iterative and time-consuming process to calculate the separation matrix. As a supporting method, principle component analysis (PCA) contributes to reduce the dimension, retrieve fast results, and disregard false sound sources. Considering the sparsity of frequency components in a mixed signal, we propose an adaptive fast speech separation algorithm based on multiple sound source localization as preprocessing to select between beamforming and frequency domain ICA according to different mixing conditions per frequency bin. First, a fast positioning algorithm allows calculating the maximum number of components per frequency bin of a mixed speech signal to prevent the occurrence of false sound sources. Then, PCA reduces the dimension to adaptively adjust the weight of beamforming and ICA for speech separation. Subsequently, the ICA separation matrix is initialized based on the sound source localization to notably reduce the iteration time and mitigate permutation ambiguity. Simulation and experimental results verify the effectiveness and speedup of the proposed algorithm.

Simultaneous Preconcentration and Separation of Native Protein Variants Using Thermal Gel Electrophoresis.

Proteins must maintain proper folding conformations and express the correct post-translational modifications (PTMs) to exhibit appropriate biological activity. However, assessing protein folding and PTMs is difficult because routine polyacrylamide gel electrophoresis (PAGE) methods lack the separation resolution necessary to identify variants of a single protein. Additionally, standard PAGE denatures proteins prior to analysis precluding determinations of folding states or PTMs. To overcome these limitations, a microfluidic thermal gel electrophoresis platform was developed to provide high-sensitivity, high-resolution analyses of native protein variants. A thermally reversible gel was utilized as a separation matrix while in its solid state (30 °C). This thermal gel provided sufficient separation resolution to identify three variants of a fluorescently labeled model protein. To increase detection sensitivity, analyte preconcentration was conducted in parallel with the separation. Continuous analyte enrichment afforded detection limits of 500 fg of protein (250 pM) while simultaneous baseline separation resolution was achieved between variants. The effects of temperature on thermal gel electrophoresis were also characterized. The unique temperature-dependent outcomes illustrated how method performance can be tuned through a thermal dimension. Ultimately, the high detection sensitivity and separation resolution provided by thermal gel electrophoresis enabled rapid screening of native protein variants.

High-Efficiency Plasma Separator Based on Immunocapture and Filtration.

The shortcomings of standard plasma-separation methods limit the point-of-care application of microfluidics in clinical facilities and at the patient’s bedside. To overcome the limitations of this inconvenient, laborious, and costly technique, a new plasma-separation technique and device were developed. This new separation method relies on immunological capture and filtration to exclude cells from plasma, and is convenient, easy to use, and cost-effective. Most of the RBCs can be captured and immobilized by antibody which coated in separation matrix, and residue cells can be totally removed from the sample by a commercially plasma purification membranes. A 400 µL anti-coagulated whole blood sample with 65% hematocrit (Hct) can be separated by the device in 5 min with only one pipette. Up to 97% of the plasma can be recovered from the raw blood sample with a separation efficiency at 100%. The recovery rate of small molecule compounds, proteins, and nucleic acid biomarkers is evaluated; there are no obvious differences from the centrifuge method. The results demonstrate that this method is an excellent replacement for traditional plasma preparation protocols.

ДОСЛІДЖЕННЯ ВПЛИВУ ГЕОМЕТРИЧНИХ РОЗМІРІВ РОБОЧОЇ ЗОНИ НА ПРОСТОРОВИЙ РОЗПОДІЛ МАГНІТНОГО ПОЛЯ В МАТРИЦІ ПОЛІГРАДІЄНТНОГО СЕПАРАТОРА

. The possibility of application of the Infolytica software package to solve the main tasks of work is substantiated. The computer 3D-model of electromagnetic separator was verified by solving of test problem. Comparison of results of magnetic induction research for characteristic points obtained during 3D modeling of electromagnetic separator with data obtained during experimental research showed a fairly high consistency. An analysis of geometric dimensions influence on distribution of magnetic field in a polygradient matrix was carried out for a random sampling containing three structural variants of an electromagnetic separator, the working air gap of which varies over a wide range. It is shown that determination of magnetic flux distribution for other structural variants of matrix that located inside sampling can be carried out by interpolation. For selected variants, three-dimensional computer models of electromagnetic separator are developed, their geometric characteristics and physical properties are described. A numerical-field analysis of magnetic field spatial distribution in working air gap of polygradient matrix of electromagnetic separator was carried out using Magnet module of Infolytica software package. For three threedimensional models of electromagnetic system of separator, numerical values of magnetic induction and magnetic flux in working air gap of polygradient matrix in center of characteristic area are obtained. The obtained results for three models are compared and variant with rational parameters of working area of separator matrix is selected.

Image-based effective feature generation for protein structural class and ligand binding prediction.

Proteins are the building blocks of all cells in both human and all living creatures of the world. Most of the work in the living organism is performed by proteins. Proteins are polymers of amino acid monomers which are biomolecules or macromolecules. The tertiary structure of protein represents the three-dimensional shape of a protein. The functions, classification and binding sites are governed by the protein’s tertiary structure. If two protein structures are alike, then the two proteins can be of the same kind implying similar structural class and ligand binding properties. In this paper, we have used the protein tertiary structure to generate effective features for applications in structural similarity to detect structural class and ligand binding. Firstly, we have analyzed the effectiveness of a group of image-based features to predict the structural class of a protein. These features are derived from the image generated by the distance matrix of the tertiary structure of a given protein. They include local binary pattern (LBP) histogram, Gabor filtered LBP histogram, separate row multiplication matrix with uniform LBP histogram, neighbor block subtraction matrix with uniform LBP histogram and atom bond. Separate row multiplication matrix and neighbor block subtraction matrix filters, as well as atom bond, are our novels. The experiments were done on a standard benchmark dataset. We have demonstrated the effectiveness of these features over a large variety of supervised machine learning algorithms. Experiments suggest support vector machines is the best performing classifier on the selected dataset using the set of features. We believe the excellent performance of Hybrid LBP in terms of accuracy would motivate the researchers and practitioners to use it to identify protein structural class. To facilitate that, a classification model using Hybrid LBP is readily available for use at http://brl.uiu.ac.bd/PL/. Protein-ligand binding is accountable for managing the tasks of biological receptors that help to cure diseases and many more. Therefore, binding prediction between protein and ligand is important for understanding a protein’s activity or to accelerate docking computations in virtual screening-based drug design. Protein-ligand binding prediction requires three-dimensional tertiary structure of the target protein to be searched for ligand binding. In this paper, we have proposed a supervised learning algorithm for predicting protein-ligand binding, which is a similarity-based clustering approach using the same set of features. Our algorithm works better than the most popular and widely used machine learning algorithms.

Development of a novel amine- and carboxyl-bifunctionalized hybrid monolithic column for non-invasive speciation analysis of chromium

A novel amine- and carboxyl-bifunctionalized organic-inorganic hybrid monolithic column (A&C-HMC) was synthesized via one-pot co-condensation of N-(2-aminoethyl)-3-aminopropyltriethoxysilane, carboxyethylsilanetriol sodium salt and tetramethoxysilane with cetyltrimethylammonium bromide and polyethylene glycol 6000 as binary porogens in this work. The introduction of the binary porogens controllably improved the morphology and pore structure of A&C-hybrid monolith (HM) and made the active sites of amine and carboxyl groups more prolific, compared with the monolith prepared with either of porogens. It is found that Cr(VI) and Cr(III) can be selectively adsorbed on A&C-HM under different pH ranges, and eluted by aqueous nitric acid solution completely. The A&C-HMC was used as needle-solid phase microextraction (SPME) matrix for direct separation and enrichment of inorganic chromium species coupled with inductively coupled plasma mass spectrometer without any oxidation/reduction treatment. Various parameters of SPME operation and analytical performance were investigated systematically, and the adsorption mechanism was also discussed and explained in depth. In view of the advantages of facile preparation, low cost, excellent speciation selectivity and high adsorption capacity to Cr(VI) and Cr(III), the A&C-HMC based SPME protocol is a promising alternative for non-disturbed speciation analysis of inorganic chromium in real environmental water samples.

Feasibility study on replacement of atmospheric distillation column with new sequences in a natural gas condensate refinery

The atmospheric distillation unit is the main separation unit in crude oil and natural gas condensate refineries. An atmospheric distillation column in a natural gas condensate unit usually consists of a single column with 3 pump-arounds and 2 side strippers. In this work a new column sequences have been synthesised for the first time using the modified separation matrix method and the results had been compared to other well-known configuration through economic, exergy and exergoeconomic analysis. The results illustrate that the conventional atmospheric distillation configuration has a medium performance compared to other sequences. As an example, comparing the results obtained from the direct sequence, to the conventional distillation sequence, leads in almost 30% less annual cost, 54% less exergy loss rate and 5% more exergoeconomic factor. However, the 84.87% exergoeconomic factor for the conventional configuration could justify the application of conventional sequence in the refinery.

Design and optimization of distillation column sequencing for NGL fractionation processes

NGL fractionation is one of the most energy-intensive facilities in hydrocarbon processing plants playing an important role in the petrochemical supply chain. Innovative schemes for multicomponent distillation processes made the opportunity to attain more justifiable fractionation processes. The current study presents a systematic procedure to design and optimize both simple and complex distillation schemes based on the separation matrix approach and applies it to a NGL feedstock. A modular design approach is presented for the ultimate rigorous evaluation. The genetic algorithm with a hybrid objective function that minimizes the TAC and checks products specifications simultaneously carries out optimization. The results demonstrate the capability of complex distillation sequences in process improvements, reducing costs and saving energy in NGL fractionation plants resulting in reductions of up to 3.5% in annual operating costs, which is very significant bearing in mind the huge NGL fractionation scale of operation.

Dynamic capture and accumulation of multiple types of magnetic particles based on fully coupled multiphysics model in multiwire matrix for high-gradient magnetic separation

High-gradient magnetic separation (HGMS) effectively separates fine weakly magnetic minerals using a magnetic matrix. The basic principle of single-wire capture of magnetic particles in HGMS has received considerable attention. In practice, however, a real matrix is made of numerous magnetic wires. Transport of magnetic particles inside a multiwire matrix under various operating conditions has not been sufficiently investigated, and it is not clear whether single-wire and multiwire matrices differ significantly. A fully coupled multiphysics model based on the idealized capture model was developed to investigate the 2D capture and accumulation of multiple types of particles in single-wire and multiwire matrices. In this model, the properties of multiple types of particles were defined. Then, particle tracing via the fluid flow model was used to calculate the dynamic capture and accumulation of particles under the determined magnetic and flow fields. The time-dependent dynamic capture mode used in this study can reveal the details of particle capture and accumulation in single-wire and multiwire matrices. All the calculations and analyses indicate that single-wire and multiwire matrices both exhibit basically the similar capture tendency as the particle size, slurry feed velocity, and magnetic induction are gradually increased, and a single-wire matrix always has a much higher capture selectivity than a multiwire matrix. This difference in selectivity between the single-wire and multiwire matrices results mainly from magnetic coupling between magnetic wires in the multiwire matrix, where the fluid flow is also quite complicated. In addition, adjacent columns of wires are staggered vertically, increasing the probability of collisions between the particles and the wires; thus, intergrowth particles that are not captured by the upstream wires are more easily captured by the downstream wires. By comparing the experimental results with the simulation results, the correctness of the HGMS recovery and grade prediction results was verified.

Karakteristike žena zavisnica od alkohola i dinamika tretmana

Objective. The main objective of the research is to determine the characteristics and dynamics of the treatment of women addicted to alcohol, at the Clinic for Addiction Diseases of the Institute for Mental Health, Belgrade. Methods. The sample consists of 136 women addicted to alcohol who were treated at the Clinic for Addiction Diseases of the Institute for Mental Health at the time of the study, as well as patients who had been treated in two wards for the last five years: Clinical ward and Daily hospital. The survey was conducted at the end of 2017 and during 2018. Secondary data sources, ie documentation from the Institute of Mental Health, were used in the research These include patient medical histories and medical records. Analysis of characteristics of women alcoholics aged 18-68 years, treated during 2018, at the Clinic for Addiction Diseases within the Institute for Mental Health in Belgrade. For the research and based on the research variables, a separate matrix was created to analyze the content which was completed for each patient covered by the research sample. Results. The age of the respondents ranges from 18 to 68 years. The results of the analysis of substance abuse in the primary family show that in more than half of the cases (51.5%) there is a history of substance abuse in the family. The key findings of the study show that the largest role in the treatment of women addiction is played by family members who participated in the treatment, and the outcome of treatment was positive with that kind of women. Conclusion. Based on the results presented and conclusions drawn to the specificity of women's alcoholism, the implication is that there is a link between family situation and support and the development of alcoholism and treatment.

The Conditions for a Linear Vibration System to Have Only Pure Rotation Modes

This paper proposes a geometric approach to the conditions for mode decoupling of a vibration system of an elastically supported single rigid body and presents the conditions that the system has only pure rotation modes of vibration. A small oscillation of a rigid body is indeed a repetitive screw motion and thus vibration modes are expressed by screws in general, which results in the difficulty involved in solving a vibration problem. The complexity of a vibration system can be alleviated for both analysis and synthesis if the system has only rotation modes. In order to acquire the decoupling techniques, this paper begins by investigating a stiffness matrix which can be separated into the sum of two rank 3 stiffness matrices, which are realizable by using co-reciprocal line vectors. From the co-reciprocity, the separable stiffness matrix can be regarded as a linear transformation between two 3-systems of screws containing only line vectors. Using the properties of the linear transformation and the screw systems, the conditions for mode decoupling, or the conditions for only pure rotation modes are derived and described by geometric relations between inertia and stiffness, and three cases of vibration systems with simple geometric nature are identified.

An Adaptive Algorithm for Calculating Crosstalk Error for Blind Source Separation

The crosstalk error is widely used to evaluate the performance of blind source separation. However, it needs to know the global separation matrix in advance, and it is not robust. In order to solve these problems, a new adaptive algorithm for calculating crosstalk error is presented, which calculates the crosstalk error by a cost function of least squares criterion, and the robustness of the crosstalk error is improved by introducing the position information of the maximum value in the global separation matrix. Finally, the method is compared with the conventional RLS algorithms in terms of performance, robustness and convergence rate. Furthermore, its validity is verified by simulation experiments and real world signals experiments.

Clustering of longitudinal interval-valued data via mixture distribution under covariance separability

ABSTRACT We consider the clustering of repeatedly measured ‘min-max’ type interval-valued data. We read the data as matrix variate data and assume the covariance matrix is separable for the model-based clustering (M-clustering). The use of a separable covariance matrix introduces several advantages in M-clustering, which include fewer samples required for a valid procedure. In addition, the numerical study shows that this structured matrix allows us to find the correct number of clusters more accurately compared to other commonly assumed covariance matrices. We apply the M-clustering with various covariance structures to clustering the longitudinal blood pressure data from the National Heart, Lung, and Blood Institute Growth and Health Study (NGHS).