Modelling High-Dimensional Data by Mixtures of Factor Analyzers

We focus on mixtures of factor analyzers from the perspective of a method for model-based density estimation from high-dimensional data, and hence for the clustering of such data. This approach enables a normal mixture model to be fitted to a sample of n data points of dimension p, where p is large relative to n. The number of free parameters is controlled through the dimension of the latent factor space. By working in this reduced space, it allows a model for each component-covariance matrix with complexity lying between that of the isotropic and full covariance structure models. We shall illustrate the use of mixtures of factor analyzers in a practical example that considers the clustering of cell lines on the basis of gene expressions from microarray experiments. (C) 2002 Elsevier Science B.V. All rights reserved.

Skin-friction measurements in high-enthalpy hypersonic boundary layers

Skin-friction measurements are reported for high-enthalpy and high-Mach-number laminar, transitional and turbulent boundary layers. The measurements were performed in a free-piston shock tunnel with air-flow Mach number, stagnation enthalpy and Reynolds numbers in the ranges of 4.4-6.7, 3-13 MJ kg(-1) and 0.16 x 10(6)-21 x 10(6), respectively. Wall temperatures were near 300 K and this resulted in ratios of wall enthalpy to flow-stagnation enthalpy in the range of 0.1-0.02. The experiments were performed using rectangular ducts. The measurements were accomplished using a new skin-friction gauge that was developed for impulse facility testing. The gauge was an acceleration compensated piezoelectric transducer and had a lowest natural frequency near 40 kHz. Turbulent skin-friction levels were measured to within a typical uncertainty of +/-7%. The systematic uncertainty in measured skin-friction coefficient was high for the tested laminar conditions; however, to within experimental uncertainty, the skin-friction and heat-transfer measurements were in agreement with the laminar theory of van Driest (1952). For predicting turbulent skin-friction coefficient, it was established that, for the range of Mach numbers and Reynolds numbers of the experiments, with cold walls and boundary layers approaching the turbulent equilibrium state, the Spalding & Chi (1964) method was the most suitable of the theories tested. It was also established that if the heat transfer rate to the wall is to be predicted, then the Spalding & Chi (1964) method should be used in conjunction with a Reynolds analogy factor near unity. If more accurate results are required, then an experimentally observed relationship between the Reynolds analogy factor and the skin-friction coefficient may be applied.

Study on the glycerolysis reaction of high free fatty acid oils for use as biodiesel feedstock

Biodiesel is the main alternative to fossil diesel and it may be produced from different feedstocks such as semi-refined vegetable oils, waste frying oils or animal fats. However, these feedstocks usually contain significant amounts of free fatty acids (FFA) that make them inadequate for the direct base catalyzed transesterification reaction (where the FFA content should be lower than 4%). The present work describes a possible method for the pre-treatment of oils with a high content of FFA (20 to 50%) by esterification with glycerol. In order to reduce the FFA content, the reaction between these FFA and an esterification agent is carried out before the transesterification reaction. The reaction kinetics was studied in terms of its main factors such astemperature, % of glycerin excess, % of catalyst used, stirring velocity and type of catalyst used. The results showed that glycerolysis is a promising pretreatment to acidic oils or fats (> 20%) as they led to the production of an intermediary material with a low content of FFA that can be used directly in thetransesterification reaction for the production of biodiesel. (C) 2011 Elsevier B.V. All rights reserved.

Derivative-free optimization and filter methods to solve nonlinear constrained problems

In real optimization problems, usually the analytical expression of the objective function is not known, nor its derivatives, or they are complex. In these cases it becomes essential to use optimization methods where the calculation of the derivatives, or the verification of their existence, is not necessary: the Direct Search Methods or Derivative-free Methods are one solution. When the problem has constraints, penalty functions are often used. Unfortunately the choice of the penalty parameters is, frequently, very difficult, because most strategies for choosing it are heuristics strategies. As an alternative to penalty function appeared the filter methods. A filter algorithm introduces a function that aggregates the constrained violations and constructs a biobjective problem. In this problem the step is accepted if it either reduces the objective function or the constrained violation. This implies that the filter methods are less parameter dependent than a penalty function. In this work, we present a new direct search method, based on simplex methods, for general constrained optimization that combines the features of the simplex method and filter methods. This method does not compute or approximate any derivatives, penalty constants or Lagrange multipliers. The basic idea of simplex filter algorithm is to construct an initial simplex and use the simplex to drive the search. We illustrate the behavior of our algorithm through some examples. The proposed methods were implemented in Java.

Derivative-free nonlinear optimization filter simplex

The filter method is a technique for solving nonlinear programming problems. The filter algorithm has two phases in each iteration. The first one reduces a measure of infeasibility, while in the second the objective function value is reduced. In real optimization problems, usually the objective function is not differentiable or its derivatives are unknown. In these cases it becomes essential to use optimization methods where the calculation of the derivatives or the verification of their existence is not necessary: direct search methods or derivative-free methods are examples of such techniques. In this work we present a new direct search method, based on simplex methods, for general constrained optimization that combines the features of simplex and filter methods. This method neither computes nor approximates derivatives, penalty constants or Lagrange multipliers.

Analysis of functionally graded sandwich plate structures with piezoelectric skins, using B-spline finite strip method

Functionally graded materials are composite materials wherein the composition of the constituent phases can vary in a smooth continuous way with a gradation which is function of its spatial coordinates. This characteristic proves to be an important issue as it can minimize abrupt variations of the material properties which are usually responsible for localized high values of stresses, and simultaneously providing an effective thermal barrier in specific applications. In the present work, it is studied the static and free vibration behaviour of functionally graded sandwich plate type structures, using B-spline finite strip element models based on different shear deformation theories. The effective properties of functionally graded materials are estimated according to Mori-Tanaka homogenization scheme. These sandwich structures can also consider the existence of outer skins of piezoelectric materials, thus achieving them adaptive characteristics. The performance of the models, are illustrated through a set of test cases. (C) 2012 Elsevier Ltd. All rights reserved.

The gambin model provides a superior fit to species abundance distributions with a single free parameter : evidence, implementation and interpretation

The species abundance distribution (SAD) has been a central focus of community ecology for over fifty years, and is currently the subject of widespread renewed interest. The gambin model has recently been proposed as a model that provides a superior fit to commonly preferred SAD models. It has also been argued that the model's single parameter (α) presents a potentially informative ecological diversity metric, because it summarises the shape of the SAD in a single number. Despite this potential, few empirical tests of the model have been undertaken, perhaps because the necessary methods and software for fitting the model have not existed. Here, we derive a maximum likelihood method to fit the model, and use it to undertake a comprehensive comparative analysis of the fit of the gambin model. The functions and computational code to fit the model are incorporated in a newly developed free-to-download R package (gambin). We test the gambin model using a variety of datasets and compare the fit of the gambin model to fits obtained using the Poisson lognormal, logseries and zero-sum multinomial distributions. We found that gambin almost universally provided a better fit to the data and that the fit was consistent for a variety of sample grain sizes. We demonstrate how α can be used to differentiate intelligibly between community structures of Azorean arthropods sampled in different land use types. We conclude that gambin presents a flexible model capable of fitting a wide variety of observed SAD data, while providing a useful index of SAD form in its single fitted parameter. As such, gambin has wide potential applicability in the study of SADs, and ecology more generally.

FMRI 3D registration based on Fourier space subsets using neural networks

In this work, we present a neural network (NN) based method designed for 3D rigid-body registration of FMRI time series, which relies on a limited number of Fourier coefficients of the images to be aligned. These coefficients, which are comprised in a small cubic neighborhood located at the first octant of a 3D Fourier space (including the DC component), are then fed into six NN during the learning stage. Each NN yields the estimates of a registration parameter. The proposed method was assessed for 3D rigid-body transformations, using DC neighborhoods of different sizes. The mean absolute registration errors are of approximately 0.030 mm in translations and 0.030 deg in rotations, for the typical motion amplitudes encountered in FMRI studies. The construction of the training set and the learning stage are fast requiring, respectively, 90 s and 1 to 12 s, depending on the number of input and hidden units of the NN. We believe that NN-based approaches to the problem of FMRI registration can be of great interest in the future. For instance, NN relying on limited K-space data (possibly in navigation echoes) can be a valid solution to the problem of prospective (in frame) FMRI registration.

The influence of experience and training in a group of novice observers: a jackknife alternative free-response receiver operating characteristic analysis

Purpose - The study evaluates the pre- and post-training lesion localisation ability of a group of novice observers. Parallels are drawn with the performance of inexperienced radiographers taking part in preliminary clinical evaluation (PCE) and ‘red-dot’ systems, operating within radiography practice. Materials and methods - Thirty-four novice observers searched 92 images for simulated lesions. Pre-training and post-training evaluations were completed following the free-response the receiver operating characteristic (FROC) method. Training consisted of observer performance methodology, the characteristics of the simulated lesions and information on lesion frequency. Jackknife alternative FROC (JAFROC) and highest rating inferred ROC analyses were performed to evaluate performance difference on lesion-based and case-based decisions. The significance level of the test was set at 0.05 to control the probability of Type I error. Results - JAFROC analysis (F(3,33) = 26.34, p < 0.0001) and highest-rating inferred ROC analysis (F(3,33) = 10.65, p = 0.0026) revealed a statistically significant difference in lesion detection performance. The JAFROC figure-of-merit was 0.563 (95% CI 0.512,0.614) pre-training and 0.677 (95% CI 0.639,0.715) post-training. Highest rating inferred ROC figure-of-merit was 0.728 (95% CI 0.701,0.755) pre-training and 0.772 (95% CI 0.750,0.793) post-training. Conclusions - This study has demonstrated that novice observer performance can improve significantly. This study design may have relevance in the assessment of inexperienced radiographers taking part in PCE or commenting scheme for trauma.

Determination of free furfuryl alcohol in foundry resins by chromatographic techniques

Two chromatographic methods, gas chromatography with flow ionization detection (GC–FID) and liquid chromatography with ultraviolet detection (LC–UV), were used to determine furfuryl alcohol in several kinds of foundry resins, after application of an optimised extraction procedure. The GC method developed gave feasibility that did not depend on resin kind. Analysis by LC was suitable just for furanic resins. The presence of interference in the phenolic resins did not allow an appropriate quantification by LC. Both methods gave accurate and precise results. Recoveries were >94%; relative standard deviations were ≤7 and ≤0.3%, respectively for GC and LC methods. Good relative deviations between the two methods were found (≤3%).

Determination of free formaldehyde in foundry resins as its 2,4-dinitrophenylhydrazone by liquid chromatography

Formaldehyde is a toxic component that is present in foundry resins. Its quantification is important to the characterisation of the resin (kind and degradation) as well as for the evaluation of free contaminants present in wastes generated by the foundry industry. The complexity of the matrices considered suggests the need for separative techniques. The method developed for the identification and quantification of formaldehyde in foundry resins is based on the determination of free carbonyl compounds by derivatization with 2,4-dinitrophenylhydrazine (DNPH), being adapted to the considered matrices using liquid chromatography (LC) with UV detection. Formaldehyde determinations in several foundry resins gave precise results. Mean recovery and R.S.D. were, respectively, >95 and 5%. Analyses by the hydroxylamine reference method gave comparable results. Results showed that hydroxylamine reference method is applicable just for a specific kind of resin, while the developed method has good performance for all studied resins.

Multiresidue method for the determination of organophosphorus pesticides in still wine and fortified wine using solid-phase microextraction and gas chromatography – Tandem mass spectrometry

A SPME-GC-MS/MS method for the determination of eight organophosphorus pesticides (azinphos-methyl, chlorpyriphos, chlorpyriphos-methyl, diazinon, fenitrothion, fenthion, malathion, and methidathion) in still and fortified wine was developed. The extraction procedure is simple, solvent free, and without any sample pretreatment. Limits of detection (LOD) and quantitation (LOQ) values in the range 0.1–14.3 lg/L and 0.2–43.3 lg/L, respectively, were obtained. The LOQ values are below the maximum residue levels (MRLs) established by European Regulation for grapes, with the exception of methidathion. Coefficients of correlation (R2) higher than 0.99 were obtained for the majority of the pesticides, in all different wines analyzed.

Chromatographic techniques for the determination of free phenol in foundry resins

Phenol is a toxic compound present in a wide variety of foundry resins. Its quantification is important for the characterization of the resins as well as for the evaluation of free contaminants present in foundry wastes. Two chromatographic methods, liquid chromatography with ultraviolet detection (LC-UV) and gas chromatography with flame ionization detection (GC-FID), for the analysis of free phenol in several foundry resins, after a simple extraction procedure (30 min), were developed. Both chromatographic methods were suitable for the determination of phenol in the studied furanic and phenolic resins, showing good selectivity, accuracy (recovery 99–100%; relative deviations <5%), and precision (coefficients of variation <6%). The used ASTM reference method was only found to be useful in the analysis of phenolic resins, while the LC and GC methods were applicable for all the studied resins. The developed methods reduce the time of analysis from 3.5 hours to about 30 min and can readily be used in routine quality control laboratories.

Quantification of endocrine disruptors and pesticides in water by gas chromatography–tandem mass spectrometry. Method validation using weighted linear regression schemes

Amulti-residue methodology based on a solid phase extraction followed by gas chromatography–tandem mass spectrometry was developed for trace analysis of 32 compounds in water matrices, including estrogens and several pesticides from different chemical families, some of them with endocrine disrupting properties. Matrix standard calibration solutions were prepared by adding known amounts of the analytes to a residue-free sample to compensate matrix-induced chromatographic response enhancement observed for certain pesticides. Validation was done mainly according to the International Conference on Harmonisation recommendations, as well as some European and American validation guidelines with specifications for pesticides analysis and/or GC–MS methodology. As the assumption of homoscedasticity was not met for analytical data, weighted least squares linear regression procedure was applied as a simple and effective way to counteract the greater influence of the greater concentrations on the fitted regression line, improving accuracy at the lower end of the calibration curve. The method was considered validated for 31 compounds after consistent evaluation of the key analytical parameters: specificity, linearity, limit of detection and quantification, range, precision, accuracy, extraction efficiency, stability and robustness.

Fractional dynamics of genetic algorithms using hexagonal space tessellation

The paper formulates a genetic algorithm that evolves two types of objects in a plane. The fitness function promotes a relationship between the objects that is optimal when some kind of interface between them occurs. Furthermore, the algorithm adopts an hexagonal tessellation of the two-dimensional space for promoting an efficient method of the neighbour modelling. The genetic algorithm produces special patterns with resemblances to those revealed in percolation phenomena or in the symbiosis found in lichens. Besides the analysis of the spacial layout, a modelling of the time evolution is performed by adopting a distance measure and the modelling in the Fourier domain in the perspective of fractional calculus. The results reveal a consistent, and easy to interpret, set of model parameters for distinct operating conditions.