A distributed EM algorithm to estimate the parameters of a finite mixture of components

In this paper, a distributed expectation maximization (DEM) algorithm is first introduced in a general form for estimating the parameters of a finite mixture of components. This algorithm is used for density estimation and clustering of data distributed over nodes of a network. Then, a distributed incremental EM algorithm (DIEM) with a higher convergence rate is proposed. After a full derivation of distributed EM algorithms, convergence of these algorithms is analyzed based on the negative free energy concept used in statistical physics. An analytical approach is also developed for evaluating the convergence rate of both incremental and distributed incremental EM algorithms. It is analytically shown that the convergence rate of DIEM is much faster than that of the DEM algorithm. Finally, simulation results approve that DIEM remarkably outperforms DEM for both synthetic and real data sets.     

Estimating the number of components in a mixture of multilayer perceptrons

Bayesian information criterion (BIC) criterion is widely used by the neural-network community for model selection tasks, although its convergence properties are not always theoretically established. In this paper we will focus on estimating the number of components in a mixture of multilayer perceptrons and proving the convergence of the BIC criterion in this frame. The penalized marginal-likelihood for mixture models and hidden Markov models introduced by Keribin [Consistent estimation of the order of mixture models, Sankhya Indian J. Stat. 62 (2000) 49–66] and, respectively, Gassiat [Likelihood ratio inequalities with applications to various mixtures, Ann. Inst. Henri Poincare 38 (2002) 897–906] is extended to mixtures of multilayer perceptrons for which a penalized-likelihood criterion is proposed. We prove its convergence under some hypothesis which involve essentially the bracketing entropy of the generalized score-function class and illustrate it by some numerical examples.     

Blind separation of a mixture of uniformly distributed source signals: A novel approach

A new, efficient algorithm for blind separation of uniformly distributed sources is proposed. The mixing matrix is assumed to be orthogonal by prewhitening the observed signals. The learning rule adaptively estimates the mixing matrix by conceptually rotating a unit hypercube so that all output signal components are contained within or on the hypercube. Under some ideal constraints, it has been theoretically shown that the algorithm is very similar to an ideal O(1/T2) convergent algorithm, which is much faster than the existing O(1/T) convergent algorithms. The algorithm has been generalized to take care of the noisy signals by adaptively dilating the hypercube in conjunction with its rotation.     

Computer-aided microfluidics (CAMF): from digital 3D-CAD models to physical structures within a day

In this paper, we introduce computer-aided microfluidics (CAMF), a process that allows the creation of complex microfluidic structures from their concept to the actual chip within a day. During design and testing of new microfluidic systems, rapid and frequent design modifications have to be carried out. For this purpose, a device using maskless projection lithography based on a digital mirror device (DMD) has been developed. Digital mask layouts may be created using any graphics program (Microsoft Paint, Adobe Photoshop) and can be used as such by the custom-written control software of the system. However, we suggest using another approach: direct importing of three-dimensional digital computer-aided design (CAD) models from which mask information can be directly parsed. This process is advantageous as commercial 3D-CAD systems allow the rapid generation of static or parameterized models which can be used for computerized analyses like, e.g., flow simulation. After model validation, the mask information is extracted from these models and directly used by the lithography device. A chip or replication master is then created by means of lithography using curable monomers or resists as, e.g., Accura 60 or SU-8. With CAMF, the whole process from digital 3D model creation to actually running the experiment can be done within a day.     

Modal: a program to calculate compositions of ternary systems within the qz-or-ab-an tetrahedron from modal-analysis data

Program MODAL transforms raw point-count data into volume and weight-percent values. By using the An-content of plagioclase and the Ab-content of alkali feldspar it reduces the percentages of the petrogenetically important feldspars to the pure components and calculates the four faces of the Qz-Or-Ab-An tetrahedron for volume-, weight- and molepercent. The program provides a routine for plotting triangular diagrams on the line printer as well as on the CALCOMP plotter. It also is possible to have the projections of points on the cotectic line plotted into the triangular diagrams.The results can be used for petrogenetic discussions and can be directly compared with other (e.g. normative) data and experimentally determined phase relationships within the granite/granodiorite system.     

On estimating the number of components in a finite mixture of power series distributions

This paper considers the problem of estimating the number of components in a finite mixture of distributions from some parametric families. An estimation procedure using a numerical algorithm for accelerating the convergence of slowly convergent sequences is developed and its asymptotic properties are investigated. The behavior of the procedure is illustrated with simulated and real data.     

Spectral mixture analysis for mapping abundance of urban surface components from the Terra/ASTER data

Using a linear unconstrained least squares (LSS) method and a non-linear artificial neural network (ANN) algorithm, we conducted a spectral mixture analysis to the Advanced Spaceborne Thermal Emission and Reflectance Radiometer (ASTER) image data in Yokohama city, Japan, for mapping the abundance of the urban surface components. ASTER is a newly developed research facility instrument. The regions of interest of four endmembers (Vegetation, Soil, High/Low albedo impervious surfaces) were determined in Maximum Noise Fraction (MNF) feature spaces. The spectral signatures of the four endmembers were then extracted from the ASTER VNIR (15-m resolution) and SWIR (30-m resolution) imagery by referring to high spatial resolution airborne imagery (The Airborne Imaging Spectrometer, AISA, with 2-m resolution) and land use/land cover map for training and testing the LSS and ANN algorithms. Experimental results indicate that ASTER VNIR and SWIR image data are capable of mapping the abundances of urban surface components with a reasonable accuracy and that the ANN outperforms the unconstrained LSS in this spectral mixture analysis.     

Superhydrophobic,passive microvalves with controllable opening threshold: exploiting plasma nanotextured microfluidics for a programmable flow switchboard

Plasma processing is used to create passive superhydrophobic on–off valves with tailored opening pressure inside microfluidic devices. First, anisotropic O₂ plasma etching on polymeric microchannels is utilized to controllably roughen (nanotexture) the bottom of the microchannel. Second, the nanotextured surfaces are hydrophobized by means of a C₄F₈ plasma deposition step through a stencil mask creating superhydrophobic stripes or patches. The superhydrophobic patches play the role of on/off valves with predesigned opening pressure threshold (in the range 40–110 mbar), determined by the microchannel dimensions and the size of the nanotexture on the patch. These valves are integrated inside microchannel networks paving the way to autonomous microfluidic devices. To this aim, we present a novel preprogrammable flow switchboard that can split and control the liquid flow for multiple analysis purposes. The proposed valves present an example of how effectively plasma nanoscience and nanotechnology can be applied to microfluidics/nanofluidics and analytical chemistry.     

Identifying hazardousness of sewer pipeline gas mixture using classification methods: a comparative study

In this work, we formulated a real-world problem related to sewer pipeline gas detection using the classification-based approaches. The primary goal of this work was to identify the hazardousness of sewer pipeline to offer safe and non-hazardous access to sewer pipeline workers so that the human fatalities, which occurs due to the toxic exposure of sewer gas components, can be avoided. The dataset acquired through laboratory tests, experiments, and various literature sources was organized to design a predictive model that was able to identify/classify hazardous and non-hazardous situation of sewer pipeline. To design such prediction model, several classification algorithms were used and their performances were evaluated and compared, both empirically and statistically, over the collected dataset. In addition, the performances of several ensemble methods were analyzed to understand the extent of improvement offered by these methods. The result of this comprehensive study showed that the instance-based learning algorithm performed better than many other algorithms such as multilayer perceptron, radial basis function network, support vector machine, reduced pruning tree. Similarly, it was observed that multi-scheme ensemble approach enhanced the performance of base predictors.

     

Estimation for the number of components in a mixture model using stepwise split-and-merge EM algorithm

The main difficulty with EM algorithm for mixture model concerns the number of components, say g. This is the question of model selection, and the EM algorithm itself could not estimate g. On the contrary, the algorithm requires g to be specified before the remaining parameters can be estimated. To solve this problem, a new algorithm, which is called stepwise split-and-merge EM (SSMEM) algorithm, is proposed. The SSMEM algorithm alternately splits and merges components, estimating g and other parameters of components simultaneously. Also, two novel criteria are introduced to efficiently select the components for split or merge. Experimental results on simulated and real data demonstrate the effectivity of the proposed algorithm.     

Gaussian mixture spline trajectory: learning from a dataset,generating trajectories without one

Most optimization-based motion planners use a naive linear initialization, which does not use previous planning experience. We present an algorithm called ‘Gaussian mixture spline trajectory’ (GMST) that leverages motion datasets for generating trajectories for new planning problems. Unlike other trajectory prediction algorithms, our method does not retrieve trajectories from a dataset. Instead, it first uses a Gaussian mixture model (GMM) to modelize the likelihood of the trajectories to be inside the dataset and then uses the GMM's parameters to generate new trajectories. As the use of the dataset is restricted only to the learning phase it can take advantage of very large datasets. Using both abstract and robot system planning problems, we show that the GMST algorithm decreases the computation time and number of iterations of optimization-based planners while increasing their success rates as compared to that obtained with linear initialization.     

Investigation of the motion of magnetic nanoparticles in microfluidics with a micro domain model

Microsystem Technologies - In this paper the magnetic nanoparticle (MNP) dynamics in a microfluidic device is investigated in the presence of an external magnetic field. The nanoparticles are used...     

Numerical simulation of novel gas separation effect in microchannel with a series of oscillating barriers

A free-molecular gas flow through the microchannel with a series of oscillating microbarriers is studied. Barriers are oscillating with high frequency in the plane, perpendicular to the axis of the channel. It is shown that probability of passing through the channel for gas molecules significantly depends on relation of oscillations frequency, molecules thermal speed and distance between barriers. Presented effect can be used for separation of gases with different molecular masses due to discrepancy of their thermal speeds. The nature of this phenomenon is studied for different values of frequencies, characteristic sizes and number of barriers. Special attention is paid to comparison of different laws of gas–surface interactions.     

Continuous review inventory models with a mixture of backorders and lost sales under fuzzy demand and different decision situations

In this paper, continuous review inventory models in which a fraction of demand is backordered and the remaining fraction is lost during the stock out period are considered under fuzzy demands. In order to find the optimal decision under different situations, two decision methods are proposed. The first one is finding a minimum value of the expected annual total cost, and the second one is maximizing the credibility of an event that the total cost in the planning periods does not exceed a certain budget level. For the first decision method, an approach of ranking fuzzy numbers by their possibilistic mean value is adopted to achieve the optimal solution. For the second one, the technique of fuzzy simulation and differential evolution algorithms are integrated to design hybrid intelligent algorithms to solve the fuzzy models. Subsequently, the two decision models are compared and some advices about inventory cash flow management are given. Further, sensitivity analysis is conducted to give more general situations to illustrate the rationality of the management advices.     

Application of the hoch-arpshofen model to ternary,quaternary, and larger systems

The Hoch-Arpshofen model is derived for ternary, quaternary and larger systems. The advantages of the model are the effect of a binary system in a ternary which is included in the basic derivation of the model, and that no special interpolation formulas are needed. Also, the maximum number of constants needed to describe a binary system is four: two for H_m, the enthalpy of mixing, and two for S_m^ex, the excess entropy of mixing.     

On the modified Reynolds equation model for the prediction of squeeze-film gas damping in a low vacuum

The Reynolds equation coupled with an effective viscosity model is often employed to predict squeeze-film damping of plate resonators in a low vacuum. Due to the lack of a sound theoretical foundation, a study is carried out to evaluate the performance of such an approach in the free-molecule regime and results are presented in this paper. An experimentally validated Monte Carlo simulation approach for the simulation of air damping is developed and employed for this study. First, effective viscosity models are developed for a parallel-plate resonator and a rotational resonator based on experimental measurements. These models are then coupled with Reynolds equation and employed to simulate air damping of resonators of the same type but with differing dimensions. The results are compared with Monte Carlo simulation results. It has been found that the modified Reynolds equation approach cannot accurately compute air damping for a general class of resonators and hence cannot serve as a predictive tool. The deficiency lies in the effective viscosity model that is assumed to be a function of Knudsen number only. Possible extensions of the modified Reynolds equation approach in the highly rarefied regime are also discussed.     

Numerical modeling of rarefied gas flow through a slit into vacuum based on the kinetic equation

A rarefied gas flow through a thin slit into vacuum is calculated on the basis of the kinetic model equations applying the discrete velocity method. The calculations are carried out for the whole range of the gas rarefaction from the free-molecular regime to the hydrodynamic one. Numerical data on the flow rate and distributions of density, bulk velocity and temperature along the symmetry axis are reported. A comparison with the corresponding results obtained previously by the direct simulation Monte Carlo method is performed. A good agreement between these results shows a reliability of the model equations which require less computational effort than the Monte Carlo method.     

A temporal-based SVM approach for the detection and identification of pollutant gases in a gas mixture

Applied Intelligence - Air toxicity and pollution phenomena are on the rise across the planet. Thus, the detection and control of gas pollution are nowadays major economic and environmental...