X-Ray Absorption Spectroscopy and Computer Modelling Study of Nanocrystalline Binary Alkaline Earth Fluorides

Nanocrystalline samples of Ba1-xCaxF2 prepared by high-energy milling show an unusually high F-ion conductivity, which exhibit a maximum in the magnitude and a minimum in the activation energy at x = 0.5. Here, we report an X-ray absorption spectroscopy (XAS) at the Ca and Sr K edges and the Ba L-3 edge and a molecular dynamics (MD) simulation study of the pure and mixed fluorides. The XAS measurements on the pure binary fluorides, CaF2, SrF2 and BaF2 show that high-energy ball-milling produces very little amorphous material, in contrast to the results for ball milled oxides. XAS measurements of Ba1-xCaxF2 reveal that for 0 < x < 1 there is considerable disorder in the local environments of the cations which is highest for x = 0.5. Hence the maximum in the conductivity corresponds to the composition with the maximum level of local disorder. The MD calculations also show a highly disordered structure consistent with the XAS results and similarly showing maximum disorder at x = 0.5.

Learning Binary Code Representations for Effective and Efficient Image Retrieval

The size of online image datasets is constantly increasing. Considering an image dataset with millions of images, image retrieval becomes a seemingly intractable problem for exhaustive similarity search algorithms. Hashing methods, which encodes high-dimensional descriptors into compact binary strings, have become very popular because of their high efficiency in search and storage capacity. In the first part, we propose a multimodal retrieval method based on latent feature models. The procedure consists of a nonparametric Bayesian framework for learning underlying semantically meaningful abstract features in a multimodal dataset, a probabilistic retrieval model that allows cross-modal queries and an extension model for relevance feedback. In the second part, we focus on supervised hashing with kernels. We describe a flexible hashing procedure that treats binary codes and pairwise semantic similarity as latent and observed variables, respectively, in a probabilistic model based on Gaussian processes for binary classification. We present a scalable inference algorithm with the sparse pseudo-input Gaussian process (SPGP) model and distributed computing. In the last part, we define an incremental hashing strategy for dynamic databases where new images are added to the databases frequently. The method is based on a two-stage classification framework using binary and multi-class SVMs. The proposed method also enforces balance in binary codes by an imbalance penalty to obtain higher quality binary codes. We learn hash functions by an efficient algorithm where the NP-hard problem of finding optimal binary codes is solved via cyclic coordinate descent and SVMs are trained in a parallelized incremental manner. For modifications like adding images from an unseen class, we propose an incremental procedure for effective and efficient updates to the previous hash functions. Experiments on three large-scale image datasets demonstrate that the incremental strategy is capable of efficiently updating hash functions to the same retrieval performance as hashing from scratch.

Calculation of sample size for stroke trials assessing functional outcome: comparison of binary and ordinal approaches

Background Many acute stroke trials have given neutral results. Sub-optimal statistical analyses may be failing to detect efficacy. Methods which take account of the ordinal nature of functional outcome data are more efficient. We compare sample size calculations for dichotomous and ordinal outcomes for use in stroke trials. Methods Data from stroke trials studying the effects of interventions known to positively or negatively alter functional outcome – Rankin Scale and Barthel Index – were assessed. Sample size was calculated using comparisons of proportions, means, medians (according to Payne), and ordinal data (according to Whitehead). The sample sizes gained from each method were compared using Friedman 2 way ANOVA. Results Fifty-five comparisons (54 173 patients) of active vs. control treatment were assessed. Estimated sample sizes differed significantly depending on the method of calculation (Po00001). The ordering of the methods showed that the ordinal method of Whitehead and comparison of means produced significantly lower sample sizes than the other methods. The ordinal data method on average reduced sample size by 28% (inter-quartile range 14–53%) compared with the comparison of proportions; however, a 22% increase in sample size was seen with the ordinal method for trials assessing thrombolysis. The comparison of medians method of Payne gave the largest sample sizes. Conclusions Choosing an ordinal rather than binary method of analysis allows most trials to be, on average, smaller by approximately 28% for a given statistical power. Smaller trial sample sizes may help by reducing time to completion, complexity, and financial expense. However, ordinal methods may not be optimal for interventions which both improve functional outcome

Use of ordinal outcomes in vascular prevention trials: comparison with binary outcomes in published trials

Background and Purpose—Vascular prevention trials mostly count “yes/no” (binary) outcome events, eg, stroke/no stroke. Analysis of ordered categorical vascular events (eg, fatal stroke/nonfatal stroke/no stroke) is clinically relevant and could be more powerful statistically. Although this is not a novel idea in the statistical community, ordinal outcomes have not been applied to stroke prevention trials in the past. Methods—Summary data on stroke, myocardial infarction, combined vascular events, and bleeding were obtained by treatment group from published vascular prevention trials. Data were analyzed using 10 statistical approaches which allow comparison of 2 ordinal or binary treatment groups. The results for each statistical test for each trial were then compared using Friedman 2-way analysis of variance with multiple comparison procedures. Results—Across 85 trials (335 305 subjects) the test results differed substantially so that approaches which used the ordinal nature of stroke events (fatal/nonfatal/no stroke) were more efficient than those which combined the data to form 2 groups (P0.0001). The most efficient tests were bootstrapping the difference in mean rank, Mann–Whitney U test, and ordinal logistic regression; 4- and 5-level data were more efficient still. Similar findings were obtained for myocardial infarction, combined vascular outcomes, and bleeding. The findings were consistent across different types, designs and sizes of trial, and for the different types of intervention. Conclusions—When analyzing vascular events from prevention trials, statistical tests which use ordered categorical data are more efficient and are more likely to yield reliable results than binary tests. This approach gives additional information on treatment effects by severity of event and will allow trials to be smaller. (Stroke. 2008;39:000-000.)

Testing the link specification in binary choice models. A semiparametric approach.

Doutoramento em Matemática.

Design of composites for thermal management: Aluminum reinforced with diamond-containing bimodal particle mixtures

A new design route is proposed in order to fabricate aluminum matrix diamond-containing composite materials with optimized values of thermal conductivity (TC) for thermal management applications. The proper size ratio and proportions of particulate diamond–diamond and diamond–SiC bimodal mixtures are selected based on calculations with predictive schemes, which combine two main issues: (i) the volume fraction of the packed particulate mixtures, and (ii) the influence of different types of particulates (with intrinsically different metal/reinforcement interfacial thermal conductances) on the overall thermal conductivity of the composite material. The calculated results are validated by comparison with measurements on composites fabricated by gas pressure infiltration of aluminum into preforms of selected compositions of particle mixtures. Despite the relatively low quality (low price) of the diamond particles used in this work, outstanding values of TC are encountered: a maximum of 770 W/m K for Al/diamond–diamond and values up to 690 W/m K for Al/diamond–SiC.

TGA/FTIR study of the pyrolysis of diammonium hydrogen phosphate–tobacco mixtures

Diammonium hydrogen phosphate (DAP) is commonly used as a flavor ingredient of commercial cigarettes. In addition, among its other uses, it is employed to expand the tobacco volume, to manufacture reconstituted tobacco sheet, and to denicotinize tobacco. However, the use of DAP as a cigarette ingredient is a controversial issue. Some authors have stated that ammonium compounds added to tobacco increase smoke ammonia and “smoke pH”, resulting in more free nicotine available in the smoke. On the other hand, other researchers have reported that the larger ammonium content of a cigarette blend due to the presence of DAP was not reflected in increased smoke ammonia. In this work, the thermal behavior of DAP, tobacco and DAP-tobacco mixtures has been studied by TGA/FTIR. The chemical processes involved in the different pyrolysis steps of DAP have been suggested. Marked changes in the pyrolytic behavior of both, tobacco and DAP have been detected when analyzing the behavior of the mixtures. A displacement of the decomposition steps mainly related to the glycerol and lignin from tobacco toward lower temperatures has been observed, whereas that associated with cellulose is displaced toward higher temperature. Additionally, no peak corresponding to the phosphorous oxides decomposition has been detected in the curves relating to the DAP-tobacco mixtures. All these features are indicative of the strong interactions between DAP and tobacco. The FTIR spectra show no significant qualitative differences between the qualitative overall composition of the gases evolved from the pyrolysis of tobacco in the absence and in the presence of DAP. Nevertheless, depending on the temperature considered, the addition of DAP contributes to a decrease in the generation of hydrocarbons and an increase in the formation of CO, CO2 and oxygenated compounds in terms of amount generated per mass of pyrolysed tobacco.

What to do when K-means clustering fails:a simple yet principled alternative algorithm

The K-means algorithm is one of the most popular clustering algorithms in current use as it is relatively fast yet simple to understand and deploy in practice. Nevertheless, its use entails certain restrictive assumptions about the data, the negative consequences of which are not always immediately apparent, as we demonstrate. While more flexible algorithms have been developed, their widespread use has been hindered by their computational and technical complexity. Motivated by these considerations, we present a flexible alternative to K-means that relaxes most of the assumptions, whilst remaining almost as fast and simple. This novel algorithm which we call MAP-DP (maximum a-posteriori Dirichlet process mixtures), is statistically rigorous as it is based on nonparametric Bayesian Dirichlet process mixture modeling. This approach allows us to overcome most of the limitations imposed by K-means. The number of clusters K is estimated from the data instead of being fixed a-priori as in K-means. In addition, while K-means is restricted to continuous data, the MAP-DP framework can be applied to many kinds of data, for example, binary, count or ordinal data. Also, it can efficiently separate outliers from the data. This additional flexibility does not incur a significant computational overhead compared to K-means with MAP-DP convergence typically achieved in the order of seconds for many practical problems. Finally, in contrast to K-means, since the algorithm is based on an underlying statistical model, the MAP-DP framework can deal with missing data and enables model testing such as cross validation in a principled way. We demonstrate the simplicity and effectiveness of this algorithm on the health informatics problem of clinical sub-typing in a cluster of diseases known as parkinsonism.

Simple approximate MAP inference for Dirichlet processes mixtures

The Dirichlet process mixture model (DPMM) is a ubiquitous, flexible Bayesian nonparametric statistical model. However, full probabilistic inference in this model is analytically intractable, so that computationally intensive techniques such as Gibbs sampling are required. As a result, DPMM-based methods, which have considerable potential, are restricted to applications in which computational resources and time for inference is plentiful. For example, they would not be practical for digital signal processing on embedded hardware, where computational resources are at a serious premium. Here, we develop a simplified yet statistically rigorous approximate maximum a-posteriori (MAP) inference algorithm for DPMMs. This algorithm is as simple as DP-means clustering, solves the MAP problem as well as Gibbs sampling, while requiring only a fraction of the computational effort. (For freely available code that implements the MAP-DP algorithm for Gaussian mixtures see http://www.maxlittle.net/.) Unlike related small variance asymptotics (SVA), our method is non-degenerate and so inherits the “rich get richer” property of the Dirichlet process. It also retains a non-degenerate closed-form likelihood which enables out-of-sample calculations and the use of standard tools such as cross-validation. We illustrate the benefits of our algorithm on a range of examples and contrast it to variational, SVA and sampling approaches from both a computational complexity perspective as well as in terms of clustering performance. We demonstrate the wide applicabiity of our approach by presenting an approximate MAP inference method for the infinite hidden Markov model whose performance contrasts favorably with a recently proposed hybrid SVA approach. Similarly, we show how our algorithm can applied to a semiparametric mixed-effects regression model where the random effects distribution is modelled using an infinite mixture model, as used in longitudinal progression modelling in population health science. Finally, we propose directions for future research on approximate MAP inference in Bayesian nonparametrics.

Activated Carbons Produced from Mixtures of Synthetic Polymers by Physical and Chemical Activation: Application on MCPA Removal

The production of AC was achieved using the most common industrial and consumer solid waste, namely PET, alone or blended with other synthetic polymer such PAN. The PET-PAN mixture (1:1 W/W %) was subjected to carbonization, with a pyrolysis yield off 31.9%, between that obtained with PET (16.9%) or PAN (42.6%) separately. By mixing PET, as a raw material, with PAN (different ratio), an improvement in the final yield of the AC production, for the same activation time, with CO2, was found.

Activated Carbons Produced from Mixtures of Synthetic Polymers by Physical and Chemical Activation: Application on MCPA Removal

The production of AC was achieved using the most common industrial and consumer solid waste, namely PET, alone or blended with other synthetic polymer such PAN. The PET-PAN mixture (1:1 W/W %) was subjected to carbonization, with a pyrolysis yield off 31.9%, between that obtained with PET (16.9%) or PAN (42.6%) separately. By mixing PET, as a raw material, with PAN (different ratio), an improvement in the final yield of the AC production, for the same activation time, with CO2, was found.

Production of activated carbons from single and mixtures of synthetic polymers

The production of activated carbons (ACs) involves two main steps: the carbonization of the carbonaceous of raw materials at temperatures below 1073 K in the absence of oxygen and the activation had realized at the temperature up to 1173 but the most useful temperature at 1073 K. In our study we used the most common industrial and consumer solid waste, namely PET, alone or blended with other synthetic polymer PAN. By mixing the two polymers in different ratios, an improvement of the yield of the AC production was found and some textural properties were enhanced by comparison with the AC prepared using each polymer separately. When all the samples were exposed through the carbonization process with a pyrolysis the mixture of PAN-PET (1:1w/w) yield around 31.9%, between that obtained with PET (16.9%) or PAN (42.6%) separately. The combine activation, with CO2 at 1073 K, allow ACs with a lower burn-off degree isothermally, when compared with those attained with PET or PAN alone, but with similarly chemicals or textural properties. The resultant ACs are microporous in their nature, as the activation time increase, the PET-PAN mixture AC are characterized by a better developed porous structure, when associated with the AC prepared from PAN. The AC prepared from PET-PAN mixture are characterized by basic surface characteristics, with a pHpzc around 10.5, which is an important characteristic for future applications on acidic pollutants removals from liquid or gaseous phase. In this study we had used the FTIR methods to determine the main functional groups in the surface of the activated carbons. The adsorbents prepared from PAN fibres presents an IR spectrum with similar characteristics to those obtained with PET wastes, but with fewer peaks and bands with less intensity, in particular for the PAN-8240 sample. This can be reflected by the stretching and deformation modes of NH bond in the range 3100 – 3300 cm-1 and 1520 – 1650 cm-1, respectively. Also, stretching mode associated to C–N, C=N, can contributed to the profile of IR spectrum around 1170 cm-1, 1585 – 1770 cm-1. And the TGA methods was used to study the loses of the precursors mass according to the excessive of the temperature. The results showed that, there were different decreasing of the mass of each precursors. PAN degradation started at almost 573 K and at 1073 K, PAN preserve more than 40% of the initial mass. PET degradation started at 650 K, but at 1073 K, it has lost 80% of the initial mass. However, the mixture of PET-PAN (1:1w/w) showed a thermogravimetric profile between the two polymers tested individually, with a final mass slightly less than 30%. From a chemical point of view, the carbonisation of PET mainly occurs in one step between 650 and 775 K.

Assessing the fit of unidimensional IRT models for binary data under model misspecification

Model misspecification affects the classical test statistics used to assess the fit of the Item Response Theory (IRT) models. Robust tests have been derived under model misspecification, as the Generalized Lagrange Multiplier and Hausman tests, but their use has not been largely explored in the IRT framework. In the first part of the thesis, we introduce the Generalized Lagrange Multiplier test to detect differential item response functioning in IRT models for binary data under model misspecification. By means of a simulation study and a real data analysis, we compare its performance with the classical Lagrange Multiplier test, computed using the Hessian and the cross-product matrix, and the Generalized Jackknife Score test. The power of these tests is computed empirically and asymptotically. The misspecifications considered are local dependence among items and non-normal distribution of the latent variable. The results highlight that, under mild model misspecification, all tests have good performance while, under strong model misspecification, the performance of the tests deteriorates. None of the tests considered show an overall superior performance than the others. In the second part of the thesis, we extend the Generalized Hausman test to detect non-normality of the latent variable distribution. To build the test, we consider a seminonparametric-IRT model, that assumes a more flexible latent variable distribution. By means of a simulation study and two real applications, we compare the performance of the Generalized Hausman test with the M2 limited information goodness-of-fit test and the Likelihood-Ratio test. Additionally, the information criteria are computed. The Generalized Hausman test has a better performance than the Likelihood-Ratio test in terms of Type I error rates and the M2 test in terms of power. The performance of the Generalized Hausman test and the information criteria deteriorates when the sample size is small and with a few items.