Sparse density estimation on the multinomial manifold

A new sparse kernel density estimator is introduced based on the minimum integrated square error criterion for the finite mixture model. Since the constraint on the mixing coefficients of the finite mixture model is on the multinomial manifold, we use the well-known Riemannian trust-region (RTR) algorithm for solving this problem. The first- and second-order Riemannian geometry of the multinomial manifold are derived and utilized in the RTR algorithm. Numerical examples are employed to demonstrate that the proposed approach is effective in constructing sparse kernel density estimators with an accuracy competitive with those of existing kernel density estimators.

Analysis and clustering of residential customers energy behavioral demand using smart meter data

Clustering methods are increasingly being applied to residential smart meter data, providing a number of important opportunities for distribution network operators (DNOs) to manage and plan the low voltage networks. Clustering has a number of potential advantages for DNOs including, identifying suitable candidates for demand response and improving energy profile modelling. However, due to the high stochasticity and irregularity of household level demand, detailed analytics are required to define appropriate attributes to cluster. In this paper we present in-depth analysis of customer smart meter data to better understand peak demand and major sources of variability in their behaviour. We find four key time periods in which the data should be analysed and use this to form relevant attributes for our clustering. We present a finite mixture model based clustering where we discover 10 distinct behaviour groups describing customers based on their demand and their variability. Finally, using an existing bootstrapping technique we show that the clustering is reliable. To the authors knowledge this is the first time in the power systems literature that the sample robustness of the clustering has been tested.

Sparse density estimation on multinomial manifold combining local component analysis

A new sparse kernel density estimator is introduced based on the minimum integrated square error criterion combining local component analysis for the finite mixture model. We start with a Parzen window estimator which has the Gaussian kernels with a common covariance matrix, the local component analysis is initially applied to find the covariance matrix using expectation maximization algorithm. Since the constraint on the mixing coefficients of a finite mixture model is on the multinomial manifold, we then use the well-known Riemannian trust-region algorithm to find the set of sparse mixing coefficients. The first and second order Riemannian geometry of the multinomial manifold are utilized in the Riemannian trust-region algorithm. Numerical examples are employed to demonstrate that the proposed approach is effective in constructing sparse kernel density estimators with competitive accuracy to existing kernel density estimators.

Using laboratory and field measurements to constrain a single habit shortwave optical parameterization for cirrus

A single habit parameterization for the shortwave optical properties of cirrus is presented. The parameterization utilizes a hollow particle geometry, with stepped internal cavities as identified in laboratory and field studies. This particular habit was chosen as both experimental and theoretical results show that the particle exhibits lower asymmetry parameters when compared to solid crystals of the same aspect ratio. The aspect ratio of the particle was varied as a function of maximum dimension, D, in order to adhere to the same physical relationships assumed in the microphysical scheme in a configuration of the Met Office atmosphere-only global model, concerning particle mass, size and effective density. Single scattering properties were then computed using T-Matrix, Ray Tracing with Diffraction on Facets (RTDF) and Ray Tracing (RT) for small, medium, and large size parameters respectively. The scattering properties were integrated over 28 particle size distributions as used in the microphysical scheme. The fits were then parameterized as simple functions of Ice Water Content (IWC) for 6 shortwave bands. The parameterization was implemented into the GA6 configuration of the Met Office Unified Model along with the current operational long-wave parameterization. The GA6 configuration is used to simulate the annual twenty-year short-wave (SW) fluxes at top-of-atmosphere (TOA) and also the temperature and humidity structure of the atmosphere. The parameterization presented here is compared against the current operational model and a more recent habit mixture model.

Inferential Implications of Over-Parametrization: A Case Study in Incomplete Categorical Data

P>In the context of either Bayesian or classical sensitivity analyses of over-parametrized models for incomplete categorical data, it is well known that prior-dependence on posterior inferences of nonidentifiable parameters or that too parsimonious over-parametrized models may lead to erroneous conclusions. Nevertheless, some authors either pay no attention to which parameters are nonidentifiable or do not appropriately account for possible prior-dependence. We review the literature on this topic and consider simple examples to emphasize that in both inferential frameworks, the subjective components can influence results in nontrivial ways, irrespectively of the sample size. Specifically, we show that prior distributions commonly regarded as slightly informative or noninformative may actually be too informative for nonidentifiable parameters, and that the choice of over-parametrized models may drastically impact the results, suggesting that a careful examination of their effects should be considered before drawing conclusions.Resume Que ce soit dans un cadre Bayesien ou classique, il est bien connu que la surparametrisation, dans les modeles pour donnees categorielles incompletes, peut conduire a des conclusions erronees. Cependant, certains auteurs persistent a negliger les problemes lies a la presence de parametres non identifies. Nous passons en revue la litterature dans ce domaine, et considerons quelques exemples surparametres simples dans lesquels les elements subjectifs influencent de facon non negligeable les resultats, independamment de la taille des echantillons. Plus precisement, nous montrons comment des a priori consideres comme peu ou non-informatifs peuvent se reveler extremement informatifs en ce qui concerne les parametres non identifies, et que le recours a des modeles surparametres peut avoir sur les conclusions finales un impact considerable. Ceci suggere un examen tres attentif de l`impact potentiel des a priori.

Bayesian analysis of extreme events with threshold estimation

The aim of this paper is to analyze extremal events using Generalized Pareto Distributions (GPD), considering explicitly the uncertainty about the threshold. Current practice empirically determines this quantity and proceeds by estimating the GPD parameters based on data beyond it, discarding all the information available be10w the threshold. We introduce a mixture model that combines a parametric form for the center and a GPD for the tail of the distributions and uses all observations for inference about the unknown parameters from both distributions, the threshold inc1uded. Prior distribution for the parameters are indirectly obtained through experts quantiles elicitation. Posterior inference is available through Markov Chain Monte Carlo (MCMC) methods. Simulations are carried out in order to analyze the performance of our proposed mode1 under a wide range of scenarios. Those scenarios approximate realistic situations found in the literature. We also apply the proposed model to a real dataset, Nasdaq 100, an index of the financiai market that presents many extreme events. Important issues such as predictive analysis and model selection are considered along with possible modeling extensions.

Ensaios sobre política monetária

Esta tese é composta de três ensaios a respeito de política monetária. O primeiro ensaio aborda o canal em que as crises financeiras aumentam a ineficiência alocativa nos países emergentes. O segundo ensaio trata do grau de não-neutralidade da moeda no Brasil de acordo com o modelo de Golosov e Lucas (2007). O terceiro ensaio estima a inclinação da hazard function da precifi cação para o Brasil pela metodologia de Finite Mixture Model.

Regressão Simplex aplicada a delineamentos de mistura e utilização do Algoritmo Boosting

In the composition of this work are present two parts. The first part contains the theory used. The second part contains the two articles. The first article examines two models of the class of generalized linear models for analyzing a mixture experiment, which studied the effect of different diets consist of fat, carbohydrate, and fiber on tumor expression in mammary glands of female rats, given by the ratio mice that had tumor expression in a particular diet. Mixture experiments are characterized by having the effect of collinearity and smaller sample size. In this sense, assuming normality for the answer to be maximized or minimized may be inadequate. Given this fact, the main characteristics of logistic regression and simplex models are addressed. The models were compared by the criteria of selection of models AIC, BIC and ICOMP, simulated envelope charts for residuals of adjusted models, odds ratios graphics and their respective confidence intervals for each mixture component. It was concluded that first article that the simplex regression model showed better quality of fit and narrowest confidence intervals for odds ratio. The second article presents the model Boosted Simplex Regression, the boosting version of the simplex regression model, as an alternative to increase the precision of confidence intervals for the odds ratio for each mixture component. For this, we used the Monte Carlo method for the construction of confidence intervals. Moreover, it is presented in an innovative way the envelope simulated chart for residuals of the adjusted model via boosting algorithm. It was concluded that the Boosted Simplex Regression model was adjusted successfully and confidence intervals for the odds ratio were accurate and lightly more precise than the its maximum likelihood version.

Quantificação e avaliação de abertura no dossel em áreas de concessões florestais: Mamuru-Arapiuns-PA

As florestas tropicais da Amazônia historicamente foram alvo de práticas pouco sustentáveis de uso da terra, restando-lhes as cicatrizes de degradação advinda da exploração madeireira predatória, do uso indiscriminado do fogo, das altas taxas de desmatamento e de outras atividades que interferem nas ações de conservação da biodiversidade desta floresta. A atuação do Estado neste cenário é necessária através de políticas que incentivem formas de uso mais sustentáveis, como é o caso das concessões florestais que buscam através do manejo florestal, contribuir para a conservação dos recursos naturais e da manutenção da biodiversidade. A geração de produtos como o Índice de Vegetação por Diferença Normalizada, Modelo Linear de Mistura Espectral e Fração de Abertura de Dossel foram realizados no intuito de criar elementos de interpretação e análise da variável abertura de dossel. Esta pesquisa teve como área de estudo a Unidade de Manejo Florestal I no Conjunto de Glebas Mamuru-Arapiuns, região oeste do estado do Pará; onde foram quantificados e avaliados a abertura de dossel nessa área de concessão florestal, através de imagens multiespectrais e fotos hemisféricas, com vistas a analisar a degradação e a qualidade do manejo executado nesta área. Os resultados obtidos mostraram que é possível estabelecer um processo de monitoramento com o uso dos sensores e técnicas aplicados, uma vez que os dados de MLME, em especial a imagem-fração solo apresentaram forte relação de covariância com os dados obtidos em campo através de fotos hemisféricas, permitindo considera-lo como uma boa ferramenta de alerta para as ações de monitoramentos das florestas amazônicas. Desta forma é possível tornar a gestão florestal mais acessível tanto ao poder público, quanto a entidades não governamentais ou privadas visando fiscalizar as ações de exploração florestal e agregar as populações que vivem nestas áreas tanto oportunidades de renda quanto a conservação florestal.

Accuracy of genomic predictions in Bos indicus (Nellore) cattle

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Emprego de imageamento passivo na análise da variabilidade espacial da turbidez no espelho d'água do Lago Paranoá, Distrito Federal

The water has an important role in human society, especially in Brazil. Its uses are multiple, including supply, energy production, recreation and others. The National Policy for Water Resources (Law No 9.433/97) states in its articles the importance of water use in accordance to their multiple uses, prioritizing the supply for humans and animals. In this approach, it is important to consider the physical and chemical quality of water to meet these demands, scope of the legal framework applied to the Brazilian water bodies according to their main uses, in order to guarantee the water quality compatible with the most demanding uses and to reduce the costs of pollution control through ongoing preventive actions. Among the various parameters that seek to analyze the physical and chemical quality of water it is intended to understand the spatial distribution of turbidity in the lake's surface, since the variation of the components that alter this parameter can be detected by means of passive remote sensing. The application of the Linear spectral mixture model allowed, satisfactorily, the identification of turbidity spatial distribution patterns in the lake.

Differential expression analysis for sequence count data via mixtures of negative binomials

The recent advent of Next-generation sequencing technologies has revolutionized the way of analyzing the genome. This innovation allows to get deeper information at a lower cost and in less time, and provides data that are discrete measurements. One of the most important applications with these data is the differential analysis, that is investigating if one gene exhibit a different expression level in correspondence of two (or more) biological conditions (such as disease states, treatments received and so on). As for the statistical analysis, the final aim will be statistical testing and for modeling these data the Negative Binomial distribution is considered the most adequate one especially because it allows for "over dispersion". However, the estimation of the dispersion parameter is a very delicate issue because few information are usually available for estimating it. Many strategies have been proposed, but they often result in procedures based on plug-in estimates, and in this thesis we show that this discrepancy between the estimation and the testing framework can lead to uncontrolled first-type errors. We propose a mixture model that allows each gene to share information with other genes that exhibit similar variability. Afterwards, three consistent statistical tests are developed for differential expression analysis. We show that the proposed method improves the sensitivity of detecting differentially expressed genes with respect to the common procedures, since it is the best one in reaching the nominal value for the first-type error, while keeping elevate power. The method is finally illustrated on prostate cancer RNA-seq data.

Corpus-Based Speech Enhancement With Uncertainty Modeling and Cepstral Smoothing

We present a new approach for corpus-based speech enhancement that significantly improves over a method published by Xiao and Nickel in 2010. Corpus-based enhancement systems do not merely filter an incoming noisy signal, but resynthesize its speech content via an inventory of pre-recorded clean signals. The goal of the procedure is to perceptually improve the sound of speech signals in background noise. The proposed new method modifies Xiao's method in four significant ways. Firstly, it employs a Gaussian mixture model (GMM) instead of a vector quantizer in the phoneme recognition front-end. Secondly, the state decoding of the recognition stage is supported with an uncertainty modeling technique. With the GMM and the uncertainty modeling it is possible to eliminate the need for noise dependent system training. Thirdly, the post-processing of the original method via sinusoidal modeling is replaced with a powerful cepstral smoothing operation. And lastly, due to the improvements of these modifications, it is possible to extend the operational bandwidth of the procedure from 4 kHz to 8 kHz. The performance of the proposed method was evaluated across different noise types and different signal-to-noise ratios. The new method was able to significantly outperform traditional methods, including the one by Xiao and Nickel, in terms of PESQ scores and other objective quality measures. Results of subjective CMOS tests over a smaller set of test samples support our claims.

Fitting of Mixtures with Unspecified Number of Components Using Cross Validation Distance Estimate

Estimation of the number of mixture components (k) is an unsolved problem. Available methods for estimation of k include bootstrapping the likelihood ratio test statistics and optimizing a variety of validity functionals such as AIC, BIC/MDL, and ICOMP. We investigate the minimization of distance between fitted mixture model and the true density as a method for estimating k. The distances considered are Kullback-Leibler (KL) and “L sub 2”. We estimate these distances using cross validation. A reliable estimate of k is obtained by voting of B estimates of k corresponding to B cross validation estimates of distance. This estimation methods with KL distance is very similar to Monte Carlo cross validated likelihood methods discussed by Smyth (2000). With focus on univariate normal mixtures, we present simulation studies that compare the cross validated distance method with AIC, BIC/MDL, and ICOMP. We also apply the cross validation estimate of distance approach along with AIC, BIC/MDL and ICOMP approach, to data from an osteoporosis drug trial in order to find groups that differentially respond to treatment.

A Pseudolikelihood Approach for Simultaneous Analysis of Array Comparative Genomic Hybridizations (aCGH)

DNA sequence copy number has been shown to be associated with cancer development and progression. Array-based Comparative Genomic Hybridization (aCGH) is a recent development that seeks to identify the copy number ratio at large numbers of markers across the genome. Due to experimental and biological variations across chromosomes and across hybridizations, current methods are limited to analyses of single chromosomes. We propose a more powerful approach that borrows strength across chromosomes and across hybridizations. We assume a Gaussian mixture model, with a hidden Markov dependence structure, and with random effects to allow for intertumoral variation, as well as intratumoral clonal variation. For ease of computation, we base estimation on a pseudolikelihood function. The method produces quantitative assessments of the likelihood of genetic alterations at each clone, along with a graphical display for simple visual interpretation. We assess the characteristics of the method through simulation studies and through analysis of a brain tumor aCGH data set. We show that the pseudolikelihood approach is superior to existing methods both in detecting small regions of copy number alteration and in accurately classifying regions of change when intratumoral clonal variation is present.