Made available in DSpace on 2014-12-17T14:08:53Z (GMT). No. of bitstreams: 1 FranciscaFNS_DISSERT.pdf: 1424383 bytes, checksum: 6d399b5a60f42e3c2b87657eb17e44e0 (MD5) Previous issue date: 2013-02-05 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior / Atualmente, ? de grande interesse o estudo de m?todos de previs?o de S?ries Temporais, ou seja, conseguir identificar e predizer algumas caracter?sticas do processo num ponto futuro. Na engenharia de petr?leo uma das atividades essenciais ? a estimativa de produ??o de ?leo existente nas reservas petrol?feras de reservat?rios maduros. O c?lculo dessas reservas ? crucial para a determina??o da viabilidade econ?mica de sua explota??o. Para tanto, a ind?stria do petr?leo faz uso de t?cnicas convencionais de modelagem de reservat?rios como simula??o num?rica matem?tica para previs?o da produ??o de petr?leo. Diante deste fato, o objetivo fundamental deste trabalho ? propor uma metodologia de An?lise de S?ries Temporais baseada nos tradicionais modelos estat?sticos de Box & Jenkins, que em conjunto com a t?cnica inteligente de Redes Neurais Artificiais (RNA s), possibilite a constru??o de um modelo h?brido de predi??o de dados de produ??o de petr?leo, tomando por base a capacidade que a rede tem em aprender com a experi?ncia e partir para generaliza??o baseada no seu conhecimento pr?vio. Para tanto, a Rede Neural ser? treinada com a finalidade de estimar e corrigir os erros associados ao modelo estat?stico de S?rie Temporal, de forma a aproximar a s?rie estimada ? s?rie de dados original. Os dados da S?rie Temporal em estudo referem-se ? curva de vaz?o de petr?leo de um reservat?rio localizado em um campo da regi?o nordeste do Brasil. A s?rie em estudo foi obtida no per?odo 31de julho do ano 1998 ate 31 de dezembro de 2007, com os dados (vaz?o) sendo obtidos com intervalos mensais, totalizando 127 meses de informa??es. O algoritmo de predi??o proposto pela Rede Neural receber? como entrada os erros gerados pelo modelo estat?stico de s?rie e fornecer? como sa?da uma estimativa do erro no tempo n+h onde h representa o horizonte de predi??o. Os erros estimados pela Rede Neural ser?o adicionados ao Modelo de S?rie Temporal com a finalidade de corrigi-lo. Por fim, ser? feito um estudo comparativo da performance preditiva do modelo de Box & Jenkins cl?ssico e o modelo de Box & Jenkins corrigido pela Rede Neural. A arquitetura recorrente em estudo neste trabalho dever? ser capaz de prover estimativas confi?veis, tanto para um horizonte de predi??o de passos simples quanto para um horizonte de m?ltiplos passos. O software utilizado para realiza??o do ajuste do modelo estat?stico de S?rie Temporal foi o R Project for Statistical Computing - vers?o 2.14.1. Para fazer as implementa??es necess?rias da Rede Neural, a ferramenta computacional utilizada foi o software Matlab Vers?o 7.0.2 (R2011a)

Este trabalho tem como objetivo introduzir os conceitos bÃsicos da Teoria das Probabilidades e apresentar noÃÃes sobre alguns modelos probabilÃsticos para o estudante do Ensino MÃdio. Iniciaremos o trabalho apresentando no capÃtulo 1 as noÃÃes de experimento determinÃstico, experimento aleatÃrio, espaÃo amostral e eventos, seguidos de algumas definiÃÃes de Probabilidade, conceitos que constituem a base para essa teoria. No capÃtulo 2 abordaremos os conceitos de Probabilidade Condicional e IndependÃncia de Eventos, apresentando alguns teoremas importantes que decorrem desses conceitos, bem como algumas de suas aplicaÃoes. No capÃtulo 3 apresentaremos de maneira simples alguns modelos probabilÃsticos discretos bastante Ãteis por modelarem de forma eficaz um bom nÃmero de experimentos aleatÃrios contribuindo assim para o cÃlculo das probabilidades de seus resultados. Por fim, no capÃtulo 4 serà apresentado o modelo probabilÃstico conhecido como DistribuiÃÃo de Poisson, que nos permite calcular a probabilidade de um evento ocorrer em um dado intervalo de tempo ou numa dada regiÃo espacial. / This work has as objective introduce the basic concepts of the Theory of Probabilities and present notions on some probabilistic models for the student of the High School. We will begin the work presented in chapter I the notions of experiment deterministic, random experiment, sample space and events, followed by some definitions of Probability concepts that constitute the basis for this theory. In chapter II we will discuss the concepts of Conditional Probability and Independence of Events showcasing some important theorems that derive from these concepts, as well as some of its applications. In chapter III we will present in a simple way some probabilistic models discrete quite useful for shape effectively a good number of random experiments thus contributing to the calculation of the probabilities of its results. Finally, in chapter IV will be presented the probability model known as Poisson distribution, which allows us to calculate the probability that an event will occur in a given time interval or in a given spatial region.

Submitted by Maria Eneide de Souza Araujo (mearaujo@ipen.br) on 2017-03-10T16:38:03Z No. of bitstreams: 0 / Made available in DSpace on 2017-03-10T16:38:03Z (GMT). No. of bitstreams: 0 / Neste trabalho foi padronizada uma solução de 14C emissor beta puro com energia máxima de 156 keV, por meio de três diferentes métodos: CIEMAT/NIST e TDCR (triple-to-double coincidence ratio) em sistemas de cintilação líquida e pelo método do traçador, em sistema de coincidências 4πβ-γ. O sistema de cintilação líquida TRICARB, equipado com dois tubos fotomultiplicadores, foi usado para aplicação do método CIEMAT/NIST, usando um padrão de 3H emissor beta puro com energia máxima de 18,7 keV como traçador de eficiência. O sistema de cintilação líquida HIDEX 300SL, equipado com três tubos fotomultiplicadores, foi utilizado para as medidas pelo método TDCR. As amostras de 14C e 3H, medidas nos sistemas de cintilação foram preparadas usando-se três coquetéis cintiladores comerciais Ultima Gold, Optiphase Hisafe3 e InstaGel-Plus a fim de comparar seu desempenho nas medidas.Todas as amostras foram preparadas com 15 mL de coquetel cintilador, em frascos de vidro com baixa concentração de potássio. Alíquotas conhecidas de solução radioativa foram depositadas nos coquetéis cintiladores. Para a variação do parâmetro indicador de quenching, foram utilizados: um carregador de nitro metano e 1 mL de água destilada. Para a padronização pelo método do traçador no sistema de coincidências 4πβ-γ, foi utilizado 60Co como emissor beta gama. As medidas foram feitas no sistema de coincidências por software SCS, usando discriminação eletrônica para alterar a eficiência beta. O comportamento da curva de extrapolação foi predito por meio do código Esquema, que utiliza a técnica de Monte Carlo. Os resultados da atividade da solução de 14C obtida pelos três métodos utilizados mostraram uma boa concordância dentro da incerteza experimental. / Dissertação (Mestrado em Tecnologia Nuclear) / IPEN/D / Instituto de Pesquisas Energeticas e Nucleares - IPEN-CNEN/SP

The Connecticut River watershed is experiencing a rapid invasion of aggressive non-native plant species, which threaten watershed function and structure. Volunteer-based monitoring programs such as the University of Massachusetts’ OutSmart Invasives Species Project, Early Detection Distribution Mapping System (EDDMapS) and the Invasive Plant Atlas of New England (IPANE) have gathered valuable invasive plant data. These programs provide a unique opportunity for researchers to model invasive plant species utilizing citizen-sourced data. This study took advantage of these large data sources to model invasive plant distribution and to determine environmental and biophysical predictors that are most influential in dispersion, and to identify a suitable presence-only model for use by conservation biologists and land managers at varying spatial scales. This research focused on the invasive plant species of high interest - Japanese stiltgrass (Mircostegium vimineum). This was identified as a threat by U.S. Fish and Wildlife Service refuge biologists and refuge managers, but for which no mutli-scale practical and systematic approach for detection, has yet been developed. Environmental and biophysical variables include factors directly affecting species physiology and locality such as annual temperatures, growing degree days, soil pH, available water supply, elevation, closeness to hydrology and roads, and NDVI. Spatial scales selected for this study include New England (regional), the Connecticut River watershed (watershed), and the U.S. Fish and Wildlife, Silvio O. Conte National Fish and Wildlife Refuge, Salmon River Division (local). At each spatial scale, three software programs were implemented: maximum entropy habitat model by means of the MaxEnt software, ecological niche factor analysis (ENFA) using Openmodeller software, and a generalized linear model (GLM) employed in the statistical software R. Results suggest that each modeling algorithm performance varies among spatial scales. The best fit modeling software designated for each scale will be useful for refuge biologists and managers in determining where to allocate resources and what areas are prone to invasion. Utilizing the regional scale results, managers will understand what areas on a broad-scale are at risk of M. vimineum invasion under current climatic variables. The watershed-scale results will be practical for protecting areas designated as most critical for ensuring the persistence of rare and endangered species and their habitats. Furthermore, the local-scale, or fine-scale, analysis will be directly useful for on-the-ground conservation efforts. Managers and biologists can use results to direct resources to areas where M. vimineum is most likely to occur to effectively improve early detection rapid response (EDRR).

A/B testing refers to the statistical procedure of experimental design and analysis to compare two treatments, A and B, applied to different testing subjects. It is widely used by technology companies such as Facebook, LinkedIn, and Netflix, to compare different algorithms, web-designs, and other online products and services. The subjects participating in these online A/B testing experiments are users who are connected in different scales of social networks. Two connected subjects are similar in terms of their social behaviors, education and financial background, and other demographic aspects. Hence, it is only natural to assume that their reactions to online products and services are related to their network adjacency. In this research, we propose to use the conditional autoregressive model (CAR) to present the network structure and include the network effects in the estimation and inference of the treatment effect. The following statistical designs are presented: D-optimal design for network A/B testing, a re-randomization experimental design approach for network A/B testing and covariate-assisted Bayesian sequential design for network A/B testing. The effectiveness of the proposed methods are shown through numerical results with synthetic networks and real social networks.

Les délétions et les duplications délétères (Variations de nombre de copies, CNV) sont identifiés dans environ 11% des individus référés dans des cliniques du neurodéveloppement pédiatrique. Certains CNVs récurrents ont été formellement associés avec des troubles du neurodéveloppement, mais la majorité des CNVs sont non-récurrents et donc trop rares pour être évalués par des études d’association. Dans cette optique, nous avons récemment développé une nouvelle approche pour estimer l’effet des CNVs non-documentés sur le quotient intellectuel non-verbal (QINV) et nous visons étendre cette approche pour l’appliquer sur une mesure de traits autistiques. Nous avons identifié les CNVs dans deux cohortes d’autisme du Simons Simplex Collection (SSC) et du MSSNG, dans leurs apparentés de premier-degré, dans une cohorte du neurodéveloppement et dans une population générale. Des modèles statistiques intégrant les scores des gènes inclus dans les CNVs ont été utilisés pour expliquer leur effet sur l’intelligence générale et sur la réciprocité sociale. Les délétions et les duplications diminuent le QINV et l’effet des duplications est 3 fois inférieur à celui des délétions. L’effet différentiel est aussi observé pour la réciprocité sociale avec un ratio d’altération de 2:1 pour les délétions et les duplications et cet effet est principalement expliqué par le QINV. Les estimés de notre modèle pour l’intelligence générale et la réciprocité sociale concordent bien avec des observations déjà publiés. Nos modèles entraînés sur des CNVs couvrant >4,500 gènes suggèrent que l’effet des CNVs sur la cognition et la réciprocité sociale est dû à leurs propriétés polygéniques. Ces modèles pourront aider dans l’interprétation des CNVs en clinique. / Deleterious deletions and duplications (copy number variations, CNVs) are identified in up to 11% of individuals referred to neurodevelopmental pediatric clinics. However, only few recurrent CNVs have been formally associated with neurodevelopmental disorders because the majority are too rare to perform individual association studies. We recently developed a new framework to estimate the effect size of undocumented CNVs on non-verbal intelligence quotient (NVIQ) and sought to extend this approach to another score measuring autistic traits. We identified CNVs in an autism sample from the Simons Simplex Collection (SSC) and MSSNG, in their first-degree relatives, in a neurodevelopmental cohort and in individuals from an unselected population. Statistical models integrating scores of the genes encompassed in the CNVs were used to explain their effect on general intelligence and on social responsiveness. Deletions and duplications decreased NVIQ and the effect of duplications was three-fold smaller than deletions. There was also a differential effect on social responsiveness: the ratio of the impairment conferred by deletions and duplications was 2:1 and this effect was mainly driven by NVIQ. Models estimates for general intelligence and social responsiveness were consistent with previously published observations. Our models, trained on CNVs encompassing >4,500 genes, suggest highly polygenic properties of CNVs with respect to cognition and social responsiveness. These models will help interpreting CNVs identified in the clinic.

Congestion on roadways and high level of uncertainty of traffic conditions are major considerations for trip planning. The purpose of this research is to investigate the characteristics and patterns of spatial and temporal correlations and also to detect other variables that affect correlation in a freeway setting. 5-minute speed aggregates from the Performance Measurement System (PeMS) database are obtained for two directions of an urban freeway – I-10 between Santa Monica and Los Angeles, California. Observations are for all non-holiday weekdays between January 1st and June 30th, 2010. Other variables include traffic flow, ramp locations, number of lanes and the level of congestion at each detector station. A weighted least squares multilinear regression model is fitted to the data; the dependent variable is Fisher Z transform of correlation coefficient. Estimated coefficients of the general regression model indicate that increasing spatial and temporal distances reduces correlations. The positive parameters of spatial and temporal distance interaction term show that the reduction rate diminishes with spatial or temporal distance. Higher congestion tends to retain higher expected value of correlation; corrections to the model due to variations in road geometry tend to be minor. The general model provides a framework for building a family of more responsive and better-fitting models for a 6.5 mile segment of the freeway during three times of day: morning, midday, and afternoon. Each model is cross-validated on two locations: the opposite direction of the freeway, and a different location on the direction used for estimation. Cross-validation results show that models are able to retain 75% or more of their original predictive capability on independent samples. Incorporation of predictor variables that describe road geometry and traffic conditions into the model works beneficially in capturing a significant portion of variance of the response. The developed regression models are thus transferrable and are apt to predict correlation on other freeway locations.

The Mississippi River Valley Alluvial Aquifer (MRVAA) is one of the most productive agricultural regions in the United States. The main objectives of this research are to identify long term trends and change points in hydrological variables (streamflow and rainfall), to assess the relationship between hydrological variables, and to evaluate the influence of global climate indices on hydrological variables. Non-parametric tests, MMK and Pettitt’s tests were used to analyze trend and change points. PCC and Streamflow elasticity analysis were used to analyze the relationship between streamflow and rainfall and the sensitivity of streamflow to rainfall changes. PCC and MLR analysis were used to evaluate the relationship between climate indices and hydrological variables and the combined effect of climate indices with hydrological variables. The results of the trend analysis indicated spatial variability within the aquifer, increase in streamflow and rainfall in the Northern region of the aquifer, while a decrease was observed in the southern region of the aquifer. Change point analysis of annual maximum, annual mean streamflow and annual precipitation revealed that statistically decreasing shifts occurred in 2001, 1998 and 1995, respectively. Results of PCC analysis indicated that streamflow and rainfall has a strong positive relationship between them with PCC values more than 0.6 in most of the locations within the basin. Results of the streamflow elasticity for the locations ranged from 0.987 to 2.33 for the various locations in the basin. Results of the PCC analysis for monthly maximum and mean streamflow showed significant maximum positive correlation coefficient for Nino 3.4. Monthly maximum rainfall showed a maximum significant positive correlation coefficient for PNA and Nino3.4 and the monthly mean rainfall showed a maximum significant positive correlation coefficient of 0.18 for Nino3.4. Results of the MLR analysis showed a maximum significant positive correlation coefficient of 0.31 for monthly maximum and mean streamflow of 0.21 and 0.23 for monthly maximum and mean rainfall, respectively. Overall, results from this research will help in understanding the impacts of global climate indices on rainfall and subsequently on streamflow discharge, so as to mitigate and manage water resource availability in the MRVAA underlying the LMRB.

Modelle und Daten sind die beiden grundlegenden Elemente in den meisten Finanzmarktstudien. Viele Arbeiten konzentrieren sich auf die Verbesserung von Modellen zur besseren Annäherung an wahre Marktmechanismen, dabei konzentriert sich ein wichtiger Teil der Literatur auf die Nutzung von Informationen aus verschiedenen Quellen. In letzter Zeit haben immer mehr Forscher die Bedeutung der Modellierung aus realen Daten erkannt, dies geht einhermit der Weiterentwicklung moderner statistischer Modelle, insbesondere dem maschinellen (statistischen) Lernen, wie z. B. rekurrente neuronale Netze, die sich in den letzten Jahren bei vielen Problemen als wirksam erwiesen haben. Es hat sich gezeigt, dass der zunehmende Trend auf innovative Datenquellen wie Textnachrichten und Satellitenbilder zuzugreifen und diese zu analysieren, sich schnell zu einer wichtigen Säule der Finanzwissenschaft entwickelt hat. Auf der anderen Seite bietet die klassische Finanzliteratur eine fundierte Basis, um die aus diesen hochentwickelten Modellen und Daten gewonnenen Ergebnisse zu hinterfragen. Basierend auf der Finanzmarktanalyse mit modernen statistischen Modellen werden in dieser Dissertation in den ersten drei Kapiteln verschiedene Themen behandelt, darunter das Portfoliomanagement in Verbindung mit Informationen aus Nachrichtennetzwerken, das Risikomanagement des aufstrebenden Bitcoin-Marktes und die Vorhersage von Zeitreihen von Stromlasten mit fortgeschrittenen statistischen Modellen. / Models and data are the two fundamental elements in most of the studies on the financial market. Many papers concentrate on improving models to better approximate the true market mechanism, while an important strand of the literature focuses on exploiting more information from various sources. Recently, more and more researchers started to realize the importance of modeling from real-world data, along with the advancement of modern statistical models, especially the machine (statistical) learning models such as Recurrent Neural Network being proved to be effective on many problems in the past few years. Hence, we saw that an uprising trend of accessing and analyzing innovative data sources, such as textual news and satellite image, has been growing fast into a major pillar in financial studies. On the other hand, the classical finance literature provides us an angle to scrutinize the results generated from those sophisticated models and data. Under the spirit of financial market analysis with modern statistical models, this dissertation is written to cover various topics, including portfolio management coupled with the information from networks of news, risk management of the emerging Bitcoin market, and electricity load time series forecasting with the advanced statistical models, in the next three chapters.

<p>In this thesis we study several inverse sampling procedures to test for homogeneity in a multivariate hypergeometric distribution. The procedures are finite population analogues of the procedures introduced in Panchapakesan et al. (1998) for the multinomial distribution. In order to develop some exact calculations for critical values not considered in Panchapakesan et al. we introduce some terminologies for target probabilities, transfer probabilities, potential target points, right intersection, and left union. Under the null and the alternative hypotheses, we give theorems to calculate the target and transfer probabilities, we then use these results to develop exact calculations for the critical values and powers of one of the procedures. We also propose a new approximate calculation. In order to speed up some of the calculations, we propose several fast algorithms for multiple summation.</p> <p>N >= 1680000, all the results are the same as those in the multinomial distribution.</p> <p>The computing results showed that the simulations agree closely with the exact results. For small population sizes the critical values and powers of the procedures are different from the corresponding multinomial procedures, but when</p> / Master of Science (MSc)