Thesis (Ph. D.)--Ohio State University, 2006. / Title from first page of PDF file. Includes bibliographical references (p. 138-148).

Thesis (Ph. D.)--Florida State University, 2006. / Advisor: Xu-Feng Niu, Florida State University, College of Arts and Sciences, Dept. of Statistics. Title and description from dissertation home page (viewed Sept. 18, 2006). Document formatted into pages; contains ix, 116 pages. Includes bibliographical references.

Universiẗat, Diss., 2004--Giessen.

Η παρούσα διατριβή μελετά λεπτομερώς, για το διάστημα 1998 – 2007, τα επίπεδα, τη δομή και τη διαχρονική εξέλιξη της φτώχειας και της ανισότητας στην Ελλάδα, καθώς και τη σχετική θέση της Ελλάδας στα πλαίσια της ΕΕ. Χρησιμοποιήθηκαν τα στοιχεία των Ερευνών Οικογενειακών Προϋπολογισμών (ΕΟΠ) του 1998/99 και 2004/05 της ΕΣΥΕ και των Ερευνών Εισοδήματος και Συνθηκών Διαβίωσης (EU SILC) των ετών 2004-2007 της EUROSTAT. Μονάδα ανάλυσης είναι το μέλος του πληθυσμού και χρησιμοποιούνται οι κατανομές της ισοδύναμης συνολικής δαπάνης και ισοδύναμου συνολικού εισοδήματος κατά κεφαλή, λαμβάνοντας υπόψη το μέγεθος και τη σύνθεση του νοικοκυριού (χρήση κλιμάκων ισοδυναμίας). Βασικό συμπέρασμα της διατριβής είναι ότι η φτώχεια και ανισότητα μειώθηκε δραματικά μεταξύ 1998 και 2004, όπου η Ελληνική κοινωνία αύξησε την ευημερία των μελών της, ιδίως αυτών που βρισκόντουσαν στο κάτω άκρο της κατανομής του εισοδήματος. Η παραπάνω ανάλυση υποδεικνύει ότι η είσοδος της Ελλάδας στο ενιαίο νόμισμα ωφέλησε κατά πολύ την Ελληνική οικονομία και το σύνολο του πληθυσμού. Το διάστημα 2004-2007 η κατάσταση φαίνεται να χειροτερεύει ελαφρά, όπου οι μικρές αυξήσεις που καταγράφονται αποτελούν τα πρώτα σημάδια της οικονομικής κρίσης που ακολούθησε μετά το τέλος της εξεταζόμενης περιόδου. Ιδιαίτερα ανησυχητικό είναι ότι η Ελλάδα βρίσκεται συνέχεια στις χειρότερες πέντε θέσεις της κατάταξης των Ευρωπαϊκών χωρών, τόσο ως προς τα επίπεδα φτώχειας όσο και της ανισότητας, ενώ, διαχρονικά χειροτερεύει η θέση της στη συνολική κατάταξη. Τα αποτελέσματα της διατριβής επιβεβαιώνουν αποτελέσματα προηγούμενων μελετών ότι η ανισότητα στην Ελλάδα πηγάζει κυρίως από τις διαφορές περισσότερο εντός παρά ανάμεσα στις πληθυσμιακές ομάδες. Αυτό οδηγεί στη διαμόρφωση πολιτικών που στόχο έχουν τη μείωση της εσωτερικής ανισότητας στις διάφορες πληθυσμιακές ομάδες, ξεκινώντας από αυτές είτε με τη μεγαλύτερη ανισότητα είτε από αυτές που συνεισφέρουν σημαντικά στη συνολική ανισότητα. Ιδιαίτερο ενδιαφέρον για τη μελέτη της φτώχειας και της ανισότητας προκαλεί η χρήση της τεχνικής bootstrap και ο στατιστικός έλεγχος των διαστημάτων εμπιστοσύνης των μέσων όρων που χρησιμοποιεί η παρούσα διατριβή για την επιβεβαίωση της σημαντικότητας των μειώσεων που παρατηρήθηκαν στους δείκτες φτώχειας και ανισότητας μεταξύ 1998 και 2004, τουλάχιστον για τρεις πληθυσμιακές ομάδες, στις οικογένειες με αρχηγό συνταξιούχο του ΟΓΑ, στις οικογένειες με μέλη στον αγροτικό τομέα και στις οικογένειες με τρία ή περισσότερα παιδιά. Η προσομοίωση σεναρίων αύξησης των ενεργειακών τιμών υπέδειξαν τη σπουδαιότητα ελέγχου των ενεργειακών τιμών, ιδιαίτερα στα καύσιμα κίνησης και θέρμανσης, καθώς ακόμα και μικρές αυξήσεις της τάξης του 10% προκαλούν μεγάλη δυσχέρεια σε αρκετά άτομα, ενώ οι πραγματικές αυξήσεις είναι σε αρκετές περιόδους πολύ μεγαλύτερες. Η προσομοίωση σεναρίων αναδιανομής, αντίστοιχα, υπέδειξε ότι υπάρχουν σενάρια αναδιανομής με μηδενικό δημοσιονομικό κόστος που δεν προβλέπεται να προκαλέσουν ιδιαίτερες κοινωνικές αναταραχές και βελτιώνουν την υπάρχουσα κατάσταση της φτώχειας και της ανισότητας. / This doctoral dissertation explores in detail, for the years 1998 – 2007, the levels, the structure and the evolution over time of poverty and inequality in Greece, and the relative position of Greece inside EE. household income and expenditure data from the last two Household Budget Surveys (HBS) conducted in 1998/99 and 2004/05 by the National Statistical Service of Greece (NSSG) and household income data from Community Statistics on Income and Living Conditions (EU SILC) survey for the years 2004 -2007 conducted by EUROSTAT are used. The individual is the unit of analysis and distributions of total equivalence expenditure and total equivalence income per capita, take into account the size and the composition of the households using “family equivalence scales”. This dissertation concludes that poverty and inequality reduced dramatically between 1998 and 2004, where Greek society and especially people in the lower edge of the distribution of income (or/and expenditure) experienced an increased prosperity. This analysis indicates that the entry in Eurozone benefited well Greek economy and the entire population. Between the years 2004 and 2007, the situation seems a little worsened, where small increments are reported THAT can be the first signs of the oncoming economic crisis that followed the end of this period. In general, Greece is constantly in the last five places of the ranking of European countries, in terms both of poverty and inequality levels, while its position worsened during that period. The dissertation results confirm results from previous studies indicating that inequality in Greece rise mainly from inequalities within rather than between population subgroups. This leads to form politics targeting the elimination of inequality within population subgroups, starting from the groups with greater internal inequality or from the groups that contribute the most in overall inequality. Of great interest in the study of poverty and inequality is the use of bootstrap technique and the statistical test of confidence intervals of means that this dissertation uses for the verification of the significance of the decrements observed in poverty and inequality indices between 1998 and 2004, at least for three population groups, the households that have at least one member with income from agricultural occupation, the households with three or more children, and the households whose head gets a pension from the Agricultural Insurance Organisation (OGA). The simulation of scenarios of increased energy prices indicates the importance of controlling energy prices, especially for gasoline and heating oil, since even small increments of 10% of their prices cause great difficulties in people with incomes around poverty line, whilst actual increments in energy prices are much larger in several periods of time. The simulation of different scenarios of income redistribution indicates, respectively, that there are redistribution scenarios with zero fiscal cost, well targeted and not expected to cause considerable social unrest.

Made available in DSpace on 2014-06-11T19:27:09Z (GMT). No. of bitstreams: 0 Previous issue date: 2013-10-25Bitstream added on 2014-06-13T20:16:06Z : No. of bitstreams: 1 alves_ej_me_rcla.pdf: 1008886 bytes, checksum: b40d9006ed7edaa96d955ce4d111dfc8 (MD5) / As técnicas de bootstrap são métodos computacionais intensivos que usam reamostragem para o cálculo de medidas de incerteza dos estimadores, tais como erros-padrão, viés e intervalos de confiança. Os métodos são aplicados a qualquer nível de modelagem e assim podem ser usados tanto na análise paramétrica quanto na não paramétrica. Os métodos bootstrap estudados são: o intervalo de confiança bootstrap padrão, o intervalo de confiança bootstrap-t, o intervalo de confiança bootstrap percentil, o intervalo de confiança bootstrap BCPB e o intervalo de confiança BCa. Utiliza-se dois códigos computacionais a serem implementados no software Matlab. Os códigos fornecem subsídios para a aplicação das técnicas estudadas a dados simulados e reais (problemas biológicos). Os intervalos de confiança bootstrap foram comparados entre si e com os métodos tradicionais de estimação da incerteza de estimadores / The bootstrap methods are intensive computational methods that use resampling to calculate measures of uncertainty of the estimators, such as standard errors, bias and confidence intervals. The methods are applicable to any level modeling and thus can be used in both parametric analysis as the non-parametric. The bootstrap methods are studied: the standard bootstrap confidence interval, the bootstrap-t confidence interval, the bootstrap percentile confidence interval, the bootstrap BCBP confidence interval and the bootstrap BCa confidence interval. Two computer codes are presented to be implemented in Matlab. Such codes allows the implementation of the techniques studied in simulated and real data (biological problems). The bootstrap confidence intervals were compared with each other and with traditional methods of estimation uncertainty estimators

Thesis title: Flexibility, Robustness and Discontinuity in Nonparametric Regression Approaches Author: Mgr. Matúš Maciak, M.Sc. Department: Department of Probability and Mathematical Statistics, Charles University in Prague Supervisor: Prof. RNDr. Marie Hušková, DrSc. huskova@karlin.mff.cuni.cz Abstract: In this thesis we focus on local polynomial estimation approaches of an unknown regression function while taking into account also some robust issues like a presence of outlying observa- tions or heavy-tailed distributions of random errors as well. We will discuss the most common method used for such settings, so called local polynomial M-smoothers and we will present the main statistical properties and asymptotic inference for this method. The M-smoothers method is especially suitable for such cases because of its natural robust flavour, which can nicely deal with outliers as well as heavy-tailed distributed random errors. Another important quality we will focus in this thesis on is a discontinuity issue where we allow for sudden changes (discontinuity points) in the unknown regression function or its derivatives respectively. We will propose a discontinuity model with different variability structures for both independent and dependent random errors while the discontinuity points will be treated in a...

This thesis presents inference for the multiple linear regression model Y = beta_1 x_1 + ... + beta_p x_p + e after model or variable selection, including prediction intervals for a future value of the response variable Y_f, and testing hypotheses with the bootstrap. If n is the sample size, most results are for n/p large, but prediction intervals are developed that may increase in average length slowly as p increases for fixed n if the model is sparse: k predictors have nonzero coefficients beta_i where n/k is large.

Resumo: A análise de confiabilidade é uma área importante, tanto para a indústria que lança novos produtos quanto para os consumidores que sempre exigem produtos cada vez melhores e que tenham uma grande durabilidade. Com isso varias técnicas foram, e ainda são, desenvolvidas para esta finalidade. Uma delas é o estimador de Kaplan-Meier, estimador da função de confiabilidade, o estimador de Kaplan-Meier é um estimador não-paramétrico assintótico, pois a assintoticidade exige que as amostras sejam grandes para que as estimativas dos valores da confiabilidade fiquem próximas aos valores reais. Uma maneira de tentar corrigir isso é a utilização de técnicas computacionalmente intensivas, tais como o método bootstrap, que é uma técnica de reamostragem proposto por Efron em 1979, que foi utilizado para avaliar a variabilidade de uma estatística qualquer. O trabalho apresenta a aplicação do método bootstrap ao estimador não-paramétrico Kaplan-Meier, e assim obter intervalos de confiança bootstrap para os valores das estimativas da confiabilidade. Foi desenvolvido um programa computacional em linguagem Fortran do método bootstrap aliado ao estimador de Kaplan-Meier, foram testadas diversas amostras de tempo de falha e comparados os resultados com os do estimador. Ao final concluiu-se que a aplicação do método resultou em resultados satisfatórios.

Submitted by ANDRE STEFANINI JIM (andre_sjim@yahoo.com.br) on 2018-09-18T19:48:38Z No. of bitstreams: 1 ANDRE_JIM-ENTREGA-SET2018 Repositório.pdf: 4995566 bytes, checksum: 08bf174134902bafb3299a10d9fc8194 (MD5) / Approved for entry into archive by Maria Lucia Martins Frederico null (mlucia@fca.unesp.br) on 2018-09-19T11:38:13Z (GMT) No. of bitstreams: 1 stefanini_aj_dr_botfca.pdf: 4995566 bytes, checksum: 08bf174134902bafb3299a10d9fc8194 (MD5) / Made available in DSpace on 2018-09-19T11:38:13Z (GMT). No. of bitstreams: 1 stefanini_aj_dr_botfca.pdf: 4995566 bytes, checksum: 08bf174134902bafb3299a10d9fc8194 (MD5) Previous issue date: 2018-05-15 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / Plantios de eucalipto podem sofrer reduções na produção de até 48% devido à ferrugem das mirtáceas causada por Autropuccinia psidii. A identificação e quantificação da doença, ou fitopatometria, e o diagnóstico precoce de doenças é fator chave no manejo florestal. Em campo, o levantamento é realizado por amostragens, que alcançam apenas 5% da área total plantada e são realizadas por equipes de avaliadores por meio do uso de escalas diagramáticas. O uso do SIG na silvicultura tem sido bem-sucedido na detecção de diversos fatores considerados favoráveis às doenças. O termo Silvicultura de Precisão consiste na adoção de ferramentas de sensoriamento remoto embarcadas em distintas plataformas juntamente com sistemas computacionais para o processamento das múltiplas informações. Dentro do campo de sensores, uma plataforma alternativa para o diagnóstico é o emprego de VANTs (Veículos Aéreos Não Tripulados) com câmeras digitais de alta resolução espacial (pixels/cm²), apresentando diversas vantagens em relação aos dados advindos de satélites. Enquanto as informações espectrais coletadas por VANTs são misturas espectrais de dosséis, o espectroradiômetro pode coletar informações de objetos, ao nível de folha, sendo assim uma importante ferramenta para comparação de resultados de análises. Diante da necessidade de se criar uma metodologia para identificação e diagnóstico de doenças em florestas, o objetivo deste trabalho foi o de associar as informações espectrais aos procedimentos estatísticos multivariados, técnicas de regressão logística e bootstrap, de modo a obter modelos de classificação consistentes quanto à identificação e separação dos níveis de severidade para a ocorrência da ferrugem das mirtáceas em plantios jovens de eucalipto. Os resultados demonstraram que a regressão logística associada à função linear discriminante alcançaram as melhores taxas de acerto para dados de laboratório (espectroradiômetro), ao passo que a análise multivariada (por componentes principais) associada à função quadrática discriminante obtiveram as melhores taxas de acerto para dados de campo (VANT). O estudo permitiu concluir que as ferramentas empregadas foram adequadas para a diferenciação de plantas sadias de plantas infectadas por A. psiidi em plantios de Eucalyptus spp. Pôde-se concluir que o uso das imagens de VANT’s oferece uma análise preliminar ampla, possibilitando determinar áreas a receberem uma investigação mais detalhada, que eventualmente, pode ser via uso de espectroradiômetro. Quanto à caracterização do patógeno, observou-se que a técnica stepwise, coincidentemente, selecionou reflectâncias próximas à atividade das clorofilas a e b, excluindo a faixa espectral do verde. / Eucalyptus plantations may suffer reductions in yield of up to 48% due to myrtle rust caused by Autropuccinia psidii. The disease identification and quantification, or phytopatometry, and the early diagnosis are a key factor in forest management. In the field, the survey is carried out by samplings, which reach only 5% of the total planted area and are carried out by teams of raters through the use of diagrammatic scales. The GIS use in forestry has been used in detecting several factors considered favorable to diseases. The term Precision Forestry consists in the adoption of remote sensing tools embedded in different platforms with computational systems for the processing of the multiple information. In terms of sensors, an alternative platform for early diagnosis is the use of UAVs equipped with high spatial resolution digital cameras (pixels / cm²), presenting several advantages compared to the satellite's data. While the spectral information collected by VANTs are canopies spectral mixtures, the spectroradiometer can collect information from objects at the leaf level and thus is an important tool for comparing results analysis. The need to create a methodology for early identification and diagnosis of forest diseases, the aim of the present work was to associate spectral information with multivariate statistical procedures, logistic regression techniques, and bootstrap, in order to obtain consistent classification models, identification of severity levels of myrtle rust in young eucalyptus plantations. The results showed that the logistic regression associated with the linear discriminant function achieved the best hit rates for laboratory data (spectroradiometer), while the multivariate analysis (by main components) associated with the discriminant quadratic function obtained the best data set rates field (UAV). We could conclude that the tools used were adequate for the differentiation of healthy plants of plants infected by A. psidii in plantations of Eucalyptus spp. It was concluded that the use of VANT's images offers a preliminary extensive analysis, making it possible to determine areas to receive a more detailed investigation, which may be via the use of a spectroradiometer. As for the characterization of the pathogen, it was observed that the stepwise technique, coincidentally, selected reflectances close to the activity of chlorophyll a and b, excluding the spectral range of green.

abstract: The Pearson and likelihood ratio statistics are well-known in goodness-of-fit testing and are commonly used for models applied to multinomial count data. When data are from a table formed by the cross-classification of a large number of variables, these goodness-of-fit statistics may have lower power and inaccurate Type I error rate due to sparseness. Pearson's statistic can be decomposed into orthogonal components associated with the marginal distributions of observed variables, and an omnibus fit statistic can be obtained as a sum of these components. When the statistic is a sum of components for lower-order marginals, it has good performance for Type I error rate and statistical power even when applied to a sparse table. In this dissertation, goodness-of-fit statistics using orthogonal components based on second- third- and fourth-order marginals were examined. If lack-of-fit is present in higher-order marginals, then a test that incorporates the higher-order marginals may have a higher power than a test that incorporates only first- and/or second-order marginals. To this end, two new statistics based on the orthogonal components of Pearson's chi-square that incorporate third- and fourth-order marginals were developed, and the Type I error, empirical power, and asymptotic power under different sparseness conditions were investigated. Individual orthogonal components as test statistics to identify lack-of-fit were also studied. The performance of individual orthogonal components to other popular lack-of-fit statistics were also compared. When the number of manifest variables becomes larger than 20, most of the statistics based on marginal distributions have limitations in terms of computer resources and CPU time. Under this problem, when the number manifest variables is larger than or equal to 20, the performance of a bootstrap based method to obtain p-values for Pearson-Fisher statistic, fit to confirmatory dichotomous variable factor analysis model, and the performance of Tollenaar and Mooijaart (2003) statistic were investigated. / Dissertation/Thesis / Doctoral Dissertation Statistics 2018