Modelo multinível é uma ferramenta estatística cada vez mais popular para análise de dados com estrutura hierárquica. O objetivo deste trabalho é propor um método para realizar a predição de observações de novos grupos usando modelos de regressão logística multinível com 2 níveis. Além disso, é apresentado e comparado dois métodos de estimação para o modelo multinível: Quase-verossimilhança Penalizada (QVP) e Quadratura de Gauss-Hermite (QGH). A idéia central está baseada no trabalho de (Jiang e Lahiri, 2006) no qual se propõe o uso do chamado melhor estimador empírico para o efeito aleatório. Através deste estimador, utilizou-se a parte fixa do modelo em conjunto com uma estimativa do desvio padrão do efeito aleatório para fazer a predição de observações de novos grupos, encontrando a probabilidade estimada dessa observação apresentar o evento de interesse, dadas suas características. / Multilevel model is an statistical tool which is becoming more and more popular in data analysis with hierachical structure. The purpose of this dissertation is to present a method to make a prediction of new group observation in multilevel logistic regression models with 2 levels. Besides, were presented and compared two estimation methods for multilevel model: Penalized Quase-likelihood and Gauss-Hermite Quadrature. The central idea is based on the paper of Jiang and Lahiri (2006), which is presented the empirical best estimator for the random effect. Through this estimator was used the fixed part of the model with an estimative of the standard deviation of the random effect to find the estimated probability of this observation presenting the target event, in accordance with its characteristic.

O milho é uma espécie alógama cujo produto comercial são os híbridos, os quais originam-se do cruzamento de duas linhagens endogâmicas. Uma forma para obtenção de tais linhagens é por meio das técnicas de indução de haploidia e posterior obtenção dos duplo-haploides, permitindo até 100% de homozigose. Essas técnicas retornam resultados importantes no melhoramento de milho. Uma variável de interesse importante, obtida a partir dessas técnicas é a taxa de indução de haploidia, a qual trata-se de uma proporção entre o número de sementes haploides e o número total de sementes. O conjunto de dados foi obtido pelo cruzamento da linhagem indutora LI- ESALQ, com cinco genótipos comerciais de milho (2B587PW, 30F53H, BM820, DKB390 e STATUS VIPTERA), em duas gerações F1 e F2, por meio de um delineamento em blocos ao acaso, na área experimental do Departamento de Genética da ESALQ/USP. A teoria dos modelos lineares generalizados (MLGs) possibilita mais opções para a distribuição da variável resposta, exigindo somente que a mesma pertença à família exponencial sob a forma canônica. Tal classe de distribuições pode ser ainda expandida para modelos que permitem efeitos aleatórios no preditor linear, caracterizando a classe dos modelos lineares generalizados mistos (MLGMs). O objetivo deste trabalho foi analisar a taxa de indução de haploidia em milho tropical, utilizando um modelo binomial misto, com efeito aleatório em nível de indivíduo. O método de estimação foi o de máxima verossimilhança. Com base em tal modelagem, verificou-se que o genótipo 30F53H, destacou-se em relação aos demais quanto à eficiência da taxa de indução de haploidia. Todas as análises foram implementadas no software R. / The maize is an allogeneic species whose commercial product are the hybrids, which are gerated by the crossing of two endogenous lines. An alternative to obtain these lines is using the haploid induction techniques and subsequent doubled haploid production, that allows up to 100% homozygous. Artificial production of doubled haploids is important in plant breeding. An important variable, that results from these techniques, is the haploid induction rate, which is a proportion between the number of haploid seeds and the total number of seeds. The data set was obtained by crossing the inductive line LI-ESALQ, with five commercial genotypes of corn (2B587PW, 30F53H, BM820, DKB390 and STATUS VIPTERA), in two generations F1 e F2, in a randomized block design, in the experimental area of Department of Genetics, ESALQ/USP. The generalized linear models (GLMs) allow more options for the variable response distribution, requiring only that it belongs to the exponential family in canonical form. The GLM class can be expanded to models that allow random effects in the linear predictor, the mixed generalized linear models (MGLM) class. This work aimed to analyze the haploid induction rate in the tropical maize. The binomial mixed model, that included random effects in individual level, was proposed. The maximum likelihood method was used to estimate the parameters. The result revealed that the genotype 30F53H stands out in relation to the others regarding the efficiency in the haploid induction rate. All the analyzes were implemented in the software R.

Submitted by Sandra Pereira (srpereira@ufrrj.br) on 2017-07-07T14:20:04Z No. of bitstreams: 1 2010 - Ana Cristina Nery de Castro.pdf: 706391 bytes, checksum: 9670ccf72b4782cc78fe6a9c8324261c (MD5) / Made available in DSpace on 2017-07-07T14:20:04Z (GMT). No. of bitstreams: 1 2010 - Ana Cristina Nery de Castro.pdf: 706391 bytes, checksum: 9670ccf72b4782cc78fe6a9c8324261c (MD5) Previous issue date: 2010-02-26 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior - CAPES / The aims of this study were to analyze the clinical signs and factors associated with the seropositivity to Brucella canis in dogs admitted to a veterinary diagnosis center in the city of Rio de Janeiro. Serum samples were collected from 841 dogs, whose participation in the study was allowed by their owners. The samples were analyzed by the Agar Gel Immunodiffusion Test. The dogs were clinically examined, whilst their owners were interviewed in order to obtain data on the animals and their raising. The associations between the explanatory variables and the infection were measured using the 2 test or Fisher exact test whenever necessary. All variables with p 0.20 at the bivariate analysis were included in the multivariable analysis by logistic regression. The backward elimination method was used to select the variables for statistical modeling. The level of significance to keep a variable in the final model was set at 5%. Seventeen (17) samples (2.0%) were reagent. The following variables were selected for the modeling: dogs with access to streets (p=0.03), ectoparasite infestation (p=0.03) tick infestation (p=0.005), reproductive activity (p=0.00001), contact with animals with reproductive problems (p=0.0009), sunlight (p=0.09), owner`s gender (p=0.16) and housing (p=0.12.In the final model, the variables tick infestation (OR= 5.47, IC 1.67- 17.86), reproductive activity (OR= 9.40, IC 2.88-30.65), contact with animals with reproductive problems (OR= 7.90, IC 2.18-28.64) and presence of sunlight inside the house (OR= 0.21, IC 0.06-0.64) remained statistically associated with seropositivity to B. canis, after confusing variables were controlled.. The clinical signs associated with the infection were abortion, scrotal dermatitis, eye alterations and presence of nodules. The results of this study suggest that canine brucellosis is a rare disease also in other similar populations, being linked to the variables associated with animal handling. This reconfirms the need of a good sanitary and reproductive management to prevent and control the disease. The detection of clinical signs associated with seropositivity to B. canis highlights the importance of a good anamnesis, as well as a detailed clinical examination of the dog upon its admittance to clinics. / Os objetivos deste estudo foram analisar os sinais cl?nicos e os fatores associados ? soropositividade para Brucella canis, em c?es atendidos num centro de diagn?stico veterin?rio do Rio de Janeiro. Amostras de soro foram coletadas de 841 de c?es, admitidos no centro de diagn?stico, cujos propriet?rios permitiram a participa??o no estudo. As amostras foram analisadas pelo exame de Imunodifus?o em Gel de Agar. Os c?es foram examinados clinicamente e uma entrevista foi realizada com seus propriet?rios para obten??o de dados sobre os animais e sua cria??o. Os testes do 2 ou exato de Fisher, quando necess?rio, foram empregados para mensurar as associa??es entre as vari?veis explicativas e a presen?a de infec??o. Todas as vari?veis com p 0,20 ? an?lise bivariada foram inclu?das na an?lise multivari?vel de regress?o log?stica. O m?todo backward elimination foi utilizado para selecionar as vari?veis para a modelagem estat?stica. O n?vel de signific?ncia para manter uma vari?vel no modelo final foi estabelecido em 5%. Foram reagentes 17 (2,0%) amostras. As vari?veis, c?es com acesso ? rua (p=0,03), infesta??o por ectoparasitas (p=0,03) e carrapatos (p=0,005), atividade reprodutiva (p=0,00001), contato com animais com dist?rbios reprodutivos (p=0,0009), luz solar (p=0,09), g?nero dos propriet?rios (p=0,16) e alojamento (p=0,12) foram selecionadas para a modelagem. No modelo final, as vari?veis infesta??o por carrapatos (OR= 5,47, IC 1,67-17,86), atividade reprodutiva (OR= 9,40, IC 2,88-30,65), contato com animais com problemas reprodutivos (OR= 7,90, IC 2,18-28,64) e presen?a de luz solar na resid?ncia (OR= 0,21, IC 0,06-0,64) se mantiveram estatisticamente associadas ? soropositividade para B. canis, ap?s o controle das vari?veis de confundimento. Os sinais cl?nicos associados ? infec??o foram o abortamento, dermatite escrotal, altera??es nos olhos e ocorr?ncia de n?dulos. Os resultados desta pesquisa indicam que a brucelose canina em popula??es semelhantes ? estudada ? uma doen?a rara e associada ?s vari?veis relativas ao manejo dos animais, refor?ando a necessidade de um bom manejo sanit?rio e reprodutivo para a preven??o e controle. A detec??o de sinais cl?nicos associados com a soropositividade para B.canis ressalta a import?ncia, de uma boa anamnese do c?o, bem como um exame cl?nico minucioso quando atendidos em cl?nicas.

Submitted by Geandra Rodrigues (geandrar@gmail.com) on 2018-07-13T15:35:50Z No. of bitstreams: 1 juvilianapereiracorrea.pdf: 44620243 bytes, checksum: 81676d4ce75e6708c8e1bcaf3ee6a27b (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2018-07-16T17:26:57Z (GMT) No. of bitstreams: 1 juvilianapereiracorrea.pdf: 44620243 bytes, checksum: 81676d4ce75e6708c8e1bcaf3ee6a27b (MD5) / Made available in DSpace on 2018-07-16T17:26:57Z (GMT). No. of bitstreams: 1 juvilianapereiracorrea.pdf: 44620243 bytes, checksum: 81676d4ce75e6708c8e1bcaf3ee6a27b (MD5) Previous issue date: 2018-06-07 / Esta dissertação contribui para os recentes debates econômicos e sociais sobre a definição do critério de elegibilidade do Programa Bolsa Família (PBF) ao verificar a probabilidade de pertencer ao público-alvo dos indivíduos inseridos no Cadastro Único até abril de 2015. A proposta foi analisar o grau de focalização do PBF. Os indicadores de focalização (IF) foram calculados mediante o critério adotado atualmente pelo Governo Federal e para os critérios alternativos estudados. Tal critério alternativo de elegibilidade terá como base uma pontuação gerada por meio de escore de propensão estimado utilizando um modelo logit, cuja variável dependente é uma dummy que indica se o domicílio pertence ao público-alvo. Como variáveis explicativas serão consideradas algumas características observáveis individuais dos cadastrados e também características relacionadas à localização das famílias e da infraestrutura do domicílio. Os resultados apontam que a grande penalização da focalização se deve principalmente a baixa cobertura do programa, ou seja, os erros de exclusão causam maior impacto que os vazamentos observados. Ao comparar os indicadores de focalização, pode-se notar que tanto para o critério de elegibilidade atual, baseado na renda declarada, quanto para o critério alternativo que os recursos destinados ao programa seriam suficientes para cobrir cerca de 70% da população pobre caso fossem destinados corretamente. Esses resultados sugerem que, apesar das críticas à forma de seleção e ao desenho institucional do programa, a utilização do critério de renda como meio de discriminar pobres de não pobres parece ser a forma mais eficiente e menos dispendiosa para um programa de grande dimensão como o Bolsa Família. / This dissertation contributes to the recent economic and social debates about the definition of the eligibility criterion of the Bolsa Familia Program (PBF) when verifying the probability of belonging to the target public of individuals enrolled in the Cadastro Unico by April 2015. The proposal was to analyze the degree of focus of the PBF. The targeting indicators were calculated using the criterion currently adopted by the Federal Government and for the alternative criteria studied. Such an alternative eligibility criterion will be based on a score generated using an estimated propensity score using a logit model, whose dependent variable is a dummy that indicates whether the household belongs to the target audience. As explanatory variables will be considered some individual observable characteristics of the registered ones and characteristics related to the location of the families and the infrastructure of the domicile. The results point out that the great penalty of the focus is mainly due to the low coverage of the program, that is, the exclusion errors cause greater impact than the observed leaks. When comparing targeting indicators, it can be noted that for both the current eligibility criterion, based on the declared income, and the alternative criterion, the resources allocated to the program would be sufficient to cover about 70% of the poor population if they were destined correctly. These results suggest that, despite criticism of the form of selection and the institutional design of the program, the use of the income criterion as a means to discriminate against the poor from the non-poor seems to be the most efficient and least expensive way for a large program such as Bolsa Familia.

Empreendedorismo tem sido objeto de incentivo no mundo e no Brasil, dada a sua significante contribuição para o desenvolvimento econômico e social de uma nação. Observa-se que a maioria das empresas, existentes ou novas, evolui de forma lenta e gradual; no entanto, reduzida parcela apresenta um padrão diferente, com crescimento elevado em faturamento ou número de colaboradores: são as empresas de alto crescimento (EACs). Tais empresas são as responsáveis por grande parte da geração de empregos (Birch, 1981; Coad, Daunfeldt, Holzl, Johansson, & Nightingale, 2014; Henrekson & Johansson, 2010; OECD, 2010). A presente tese procurou identificar as variáveis que ajudam a explicar o desempenho das EACs no Brasil, classificadas conforme critério da OECD (2007). Foi conduzida uma pesquisa com 470 empresas brasileiras, que coletou mais de 30 variáveis preditoras categóricas ou métricas, utilizadas no modelo Regressão Logística. Foram identificadas algumas variáveis alinhadas com a literatura e outras menos intuitivas e documentadas. Comprovou-se o aumento da probabilidade de EACs quando se relacionavam com aceleradoras, recebiam premiações ou eram spin-offs de outras empresas. Em função das altas taxas de juros bancários e da cultura empreendedora no Brasil, surpreendeu o impacto positivo de empréstimos bancários e a percepção dos empreendedores sobre registrar marcas comerciais, bem como o impacto negativo da percepção sobre propaganda em mídia digital e doações de instituições de fomento, relacionadas às EACs. Análises adicionais com o subgrupo de EACs caracterizadas como gazelas foram feitas. Embora apresente limitações de surveys e outras, a tese confirmou parte dos resultados da literatura sobre empreendedorismo e identificou avenidas para futuras pesquisas. / Entrepreneurship has been object of encouragement in the world and in Brazil, given its significant contribution to the economic and social development of a nation. It is observed that the majority of companies, existing or new, are developing slowly and gradually; however, small share presents a different pattern, with high growth in sales or number of employees: they are the high growth firms (HGFs). These firms are responsible for a large part of job creation (Birch, 1981; Coad, Daunfeldt, Holzl, Johansson, & Nightingale, 2014, Henrekson & Johansson, 2010, OECD, 2010). This thesis aimed to identify the variables that help to explain the performance of HGFs in Brazil, according to OECD (2007) criterion. A survey with 470 Brazilian companies was conducted, collecting more than 30 categorical or metric predictor variables, used in the Logistic Regression model. Some identified variables were aligned to literature, but others less intuitive or documented. It was confirmed the increase in the probability of HGFs when they related to accelerators, received awards, or were spin-offs of other companies. As a consequence of the high banking interest rates and the entrepreneurship culture in Brazil, surprised the positive impact of bank loans and the entrepreneurs\' perception of trademark registration, as well as the negative impact of perception on advertising in digital media and donations from development institutions, related to HGFs. Additional analyzes with the subgroup of HGFs characterized as gazelles were made. Although it presents limitations of surveys and others, the thesis confirmed part of the results of the literature on entrepreneurship and identified avenues for future researches.

O efeito disposição, originalmente proposto por Shefrin e Statman (1985), preconiza que os investidores tendem a vender ações com lucro em um curto período de tempo e manter ações com prejuízo por um longo período de tempo. A despeito da ampla gama de evidências sobre o assunto, as razões que levariam os investidores a manifestar esse viés comportamental ainda é motivo de uma controvérsia importante entre motivações racionais e comportamentais. Neste trabalho, o objetivo foi testar duas motivações comportamentais concorrentes para explicar o efeito disposição: a teoria perspectiva e o viés da reversão à média. Para cumprir esse objetivo, foi feita uma análise das transações mensais de compra e venda de uma amostra de 51 fundos de investimento em ações brasileiros, no período de 2002 a 2008. A análise envolveu a estimação de dois modelos de regressão de variável dependente qualitativa. O primeiro consistiu em um modelo logit binário cujo propósito foi determinar a probabilidade de um gestor realizar um ganho ou uma perda de capital em razão de variáveis de retorno das ações. O segundo foi um modelo logit ordenado cujo objetivo foi verificar a existência de uma relação entre as variáveis de retorno e o volume monetário vendido das ações. Em ambos os modelos, os parâmetros estimados para as variáveis de retorno das ações foram interpretados como um coeficiente de disposição, sendo que a proposição desse coeficiente consistiu na principal contribuição da pesquisa. Os resultados dos modelos estimados trouxeram evidências de que a teoria perspectiva parece permear o processo decisório dos gestores dos fundos analisados. Já no caso da hipótese de que o efeito disposição é decorrente do viés da reversão à média, não foi possível corroborá-la com base nos resultados aqui relatados. / The disposition effect, originally proposed by Shefrin and Statman (1985), predicts that investors tend to sell winning stocks too soon and ride losing stocks too long. Despite the wide range of research evidence about this issue, the reasons that lead investors to act this way is still subject to much controversy between rational and behavioral explanations. In this thesis, the main goal was to test two competing behavioral motivations to justify the disposition effect: prospect theory and mean reversion bias. To achieve this goal, an analysis of monthly transactions for a sample of 51 Brazilian stock funds from 2002 to 2008 was conducted. The analysis involved the estimation of two regression models with qualitative dependent variable. The first one consisted of a binary logit model whose purpose was to set the probability of a manager to realize a capital gain or loss as a function of the stock return. The second one was an ordered logit model whose objective was to verify the existence of a relationship between stock returns and the monetary volume sold. In both models, the estimated parameters for the stock return variables were interpreted as a disposition coefficient and the proposition of this coefficient was the main contribution of the research. The results of the estimated models brought evidence that prospect theory seems to guide the decision making process of the managers of the analyzed funds. The hypothesis that the disposition effect is due to mean reversion bias could not be confirmed based on the results reported here.

Klarna uses a logistic regression to estimate the probability that an e-store customer will default on its given credit. The logistic regression is a linear statistical model which cannot detect non-linearities in the data. The aim of this project has been to develop a program which can be used to find suitable non-linear interaction-variables. This can be achieved using a Restricted Boltzmann Machine, an unsupervised neural network, whose hidden nodes can be used to model the distribution of the data. By using the hidden nodes as new variables in the logistic regression it is possible to see which nodes that have the greatest impact on the probability of default estimates. The contents of the hidden nodes, corresponding to different parts of the data distribution, can be used to find suitable interaction-variables which will allow the modelling of non-linearities. It was possible to find the data distribution using the Restricted Boltzmann Machine and adding its hidden nodes to the logistic regression improved the model's ability to predict the probability of default. The hidden nodes could be used to create interaction-variables which improve Klarna's internal models used for credit risk estimates. / Klarna använder en logistisk regression för att estimera sannolikheten att en e-handelskund inte kommer att betala sina fakturor efter att ha givits kredit. Den logistiska regressionen är en linjär modell och kan därför inte upptäcka icke-linjäriteter i datan. Målet med detta projekt har varit att utveckla ett program som kan användas för att hitta lämpliga icke-linjära interaktionsvariabler. Genom att införa dessa i den logistiska regressionen blir det möjligt att upptäcka icke-linjäriteter i datan och därmed förbättra sannolikhetsestimaten. Det utvecklade programmet använder Restricted Boltzmann Machines, en typ av oövervakat neuralt nätverk, vars dolda noder kan användas för att hitta datans distribution. Genom att använda de dolda noderna i den logistiska regressionen är det möjligt att se vilka delar av distributionen som är viktigast i sannolikhetsestimaten. Innehållet i de dolda noderna, som motsvarar olika delar av datadistributionen, kan användas för att hitta lämpliga interaktionsvariabler. Det var möjligt att hitta datans distribution genom att använda en Restricted Boltzmann Machine och dess dolda noder förbättrade sannolikhetsestimaten från den logistiska regressionen. De dolda noderna kunde användas för att skapa interaktionsvariabler som förbättrar Klarnas interna kreditriskmodeller.

The persistence rate in online classes remains significantly lower than in comparable face-to-face classes. There is a lack of research on strategies to help learners persist in online classes. The purpose of this study was to investigate the relationship between face-to-face orientations, instructor verbal immediacy behaviors, and persistence in online classes. It also explored the relationship between verbal immediacy and student satisfaction. Due to the disagreement in the literature, this study also investigated the relationship between student demographic variables, grade point average, and persistence. This quantitative study combined both causal-comparative and correlational research methods. The study took place at a community college in the southeastern United States. The participants were 171 students enrolled in online classes selected for inclusion in the study using a purposive sampling method. Students were invited to attend a face-to-face orientation prior to the start of the semester. Prior to the end of the course, students were invited to complete a survey consisting of modified versions of Gorham’s (1988) Verbal Immediacy Scale and the college’s Student Perception of Instruction survey. At the end of the semester, demographic and course completion data were collected from the college’s student information system. A chi-square analysis revealed that there was not a significant relationship between orientation attendance and persistence. A Pearson correlation analysis and a linear regression analysis revealed that there was not a significant relationship between instructor verbal immediacy and persistence. A second linear regression analysis revealed a significant relationship between verbal immediacy and student satisfaction. Perceived instructor verbal immediacy explained 33% of the variance in satisfaction. A logistic regression analysis found that while GPA was a significant predictor of persistence, age, gender, and race/ethnicity were not. The results of this study suggest that online faculty should receive training on communication strategies, such as verbal immediacy behaviors. Colleges should also consider using predictor variables, such as GPA, to identify and assist at-risk students. This study was one of the first to examine the relationship between verbal immediacy and persistence in online classes. Future research should continue to investigate this topic as well as other strategies to help students persist.

The two main causes of packet corruption in wireless sensor network are multipath fading and WiFi interference. Identifying the cause of the corruption can be used to improve the reliability of the transmission. If the corruption is caused by WiFi interference, the network could change the channel to a free one. If it is caused by multipath fading, the network could reroute the traffic away from the obstacles or shorten the distance. This project proposes a new method of corruption-causes-identification for the two causes mentioned. It is an immediate online diagnostic tool for IEEE 802.15.4 packets with a retransmission mechanism. It provides a statistical boundary with a processed deviation of RSSI value and the frame symbol error rate, and also a rechecking mechanism to check the decisions. In this model, the deviation of RSSI value is measured by the estimated RSSI value of correct packets and the current detected RSSI value. The benefit of the deviation design is that the statistical model hardly needs to retrain and correct the parameters in different transmission environments. The project also discusses two rechecking mechanism methods, one employs an individual secondary classification with its own characters; the other combines the neighbor packets' features to smooththe probable errors. To investigate the performance of the "frame symbol errorrate and deviation of RSSI values"-based diagnostic tool, the evaluation parts provide a comparison of different length packets. The conclusion is that this diagnostic tool provides a correct judgment of the accuracy of packet corruption caused by multipath fading up to 98.70%, and an accuracy of up to 92.99% for the WiFi-interferenced packet corruption.

A long record of daily data is used to study the lake levels of Lake Neusiedl, a large steppe lake at the eastern border of Austria. Daily lake level changes are modeled as functions of precipitation, temperature, and wind conditions. The occurrence and the amount of daily precipitation are modeled with logistic regressions and generalized linear models.