Rozprawy doktorskie na temat „Multiple Regression Estimate”

Nutrients are of critical concern in water quality assessment. The development of empirical models to estimate mean nutrient concentrations, based on satellite derived land use, could aid water resource managers. Models using land use acreages outperformed those using percentages, and discrete urban land uses were superior to lumped urban. Regressions of the combinations of two, three and four of the eight possible land use variables were investigated. Sensitivity analyses, with one stream deleted each series, identified robust combinations of variables at each level. Although uncertainty exists regarding the final regression coefficients, five of the six actual measured nitrate and total phosphorus mean concentrations were within the 95 percent confidence limits.

Market multiples are a common and simple tool for estimation of corporate value. It can express temporal dynamics and differences in markets, industries and firms. Despite their practical usefulness, some critical problems remains which continue to be debated. This thesis investigates if there exists characteristics for explaining market capitalization by market multiples within the life science industry in Sweden. The approach follows well known theory of multiple linear regression analysis. The results indicated only a linear relationship between the market cap and the R\&D expenditures of a company. This does not mean that the other explanatory variables does not have effect on market cap only that there is no linear relationship that could be statistically proven.
Värderingsmultiplar är ett vanligt och enkelt verktyg för att approximera företags värde. Det kan beskriva temporär dynamik och skillnader hos marknader, industrier och bolag. Trots dess praktiska användbarhet finns en del kritiska problem som fortfarande debatteras. Denna uppsats undersöker om det existerar några egenskaper för att förklara marknadsvärdet med hjälp av värderingsmultiplar inom life science industrin i Sverige. Tillvägagångssättet följer välkänd teori om multipel linjär regressions analys. Resultaten visade att det endast finns ett samband mellan marknadsvärdet och utgifter för forskning och utveckling för ett bolag. Detta innebär inte att de andra variablerna inte har någon effekt på marknadsvärdet, utan att det inte finns ett linjärt samband som kan bevisas på ett statistiskt vis.

As the impacts of human activities increase sediment transport by aquatic systems the need to accurately quantify this transport becomes paramount. Turbidity is recognized as an effective tool for monitoring suspended sediments in aquatic systems, and with recent technological advances turbidity can be measured in-situ remotely, continuously, and at much finer temporal scales than was previously possible. Although turbidity provides an improved method for estimation of suspended-sediment concentration (SSC), compared to traditional discharge-based methods, there is still significant variability in turbidity-based SSC estimates and in sediment loadings calculated from those estimates. The purpose of this study was to improve the turbidity-based estimation of SSC. Working at two monitoring sites on the Roanoke River in southwestern Virginia, stage, turbidity, and other water-quality parameters and were monitored with in-situ instrumentation, suspended sediments were sampled manually during elevated turbidity events; those samples were analyzed for SSC and for physical properties; rainfall was quantified by geologic source area. The study identified physical properties of the suspended-sediment samples that contribute to SSC-estimation variance and hydrologic variables that contribute to variance in those physical properties. Results indicated that the inclusion of any of the measured physical properties, which included grain-size distributions, specific surface-area, and organic carbon, in turbidity-based SSC estimation models reduces unexplained variance. Further, the use of hydrologic variables, which were measured remotely and on the same temporal scale as turbidity, to represent these physical properties, resulted in a model which was equally as capable of predicting SSC. A square-root transformed turbidity-based SSC estimation model developed for the Roanoke River at Route 117 monitoring station, which included a water level variable, provided 63% less unexplained variance in SSC estimations and 50% narrower 95% prediction intervals for an annual loading estimate, when compared to a simple linear regression using a logarithmic transformation of the response and regressor (turbidity). Unexplained variance and prediction interval width were also reduced using this approach at a second monitoring site, Roanoke River at Thirteenth Street Bridge; the log-based transformation of SSC and regressors was found to be most appropriate at this monitoring station. Furthermore, this study demonstrated the potential for a single model, generated from a pooled set of data from the two monitoring sites, to estimate SSC with less variance than a model generated only from data collected at this single site. When applied at suitable locations, the use of this pooled model approach could provide many benefits to monitoring programs, such as developing SSC-estimation models for multiple sites which individually do not have enough data to generate a robust model or extending the model to monitoring sites between those for which the model was developed and significantly reducing sampling costs for intensive monitoring programs.
Master of Science

Submitted by Maike Costa (maiksebas@gmail.com) on 2016-04-25T12:59:29Z No. of bitstreams: 1 arquivo total.pdf: 2399944 bytes, checksum: b31ffd454938b9e66e2494032cb6d9ee (MD5)
Made available in DSpace on 2016-04-25T12:59:29Z (GMT). No. of bitstreams: 1 arquivo total.pdf: 2399944 bytes, checksum: b31ffd454938b9e66e2494032cb6d9ee (MD5) Previous issue date: 2015-08-28
In recent years, several studies where published regarding to the estimation of variables related to the use of electricity, where the most varied methodologies are used to perform modeling and estimation of demand for energy of countries, States, companies in general and educational systems. In this dissertation where chosen this last category and the focus is on Higher Education Institutions (HEIs). Looking for drawing up an estimate of Wing Maxim Demand (WMD), monthly of electrical energy power, for the (HEIs), from the amount of students and, if necessary, from other causal variables, which can contribute to managerial way for the renegotiation of contracts with concessionaires that lead to annual cost savings and still contribute to a better control of the levels of maxim demand of electricity. To achieve this objective, it was realized a review of the literature regarding to the variables that could introduce correlation with the dependent variable WMD. This review indicated several methodologies that could contribute to the solution of the problem proposed: Markov Chain, Support vector Regression methodology, Genetic Programming Model and Artificial Neural Networks. It was adopted the methodology of Multiple Linear Regression (MLR) because it is less complex and a methodology directed at large companies. It was selected an IES and were carried out interviews with some engineers and technician of his electrical engineering division, seeking to better understand energy use and the behavior of the variable WMD in this IES being made available the reports of power energy monitoring where the WMD data of January-December 2008 of 2014 were contained. So on the basis of these data and documental research of the independent variables, and, through the methodologies of Multiple Linear Regression (MLR), it was developed a model from the data of 72 months which had their waste evaluated, showing a coefficient of determination R ^ 2 equal to 0.883. Independent variables that remained in the model, from the use of the backward method, were 4 (four) Dummy variables associated with the years, six variables of this type associated with the months and a variable which is the product of school days for graduates and the quantity of graduate students registered. This model was able to identify seasonality presents in the behavior of the WMD of this HIE. It would allow the hiring of WMD per month, that would make savings of 57% compared to the traditional contracting mode (WMD fixed for the entire period), considering the period from July to December, before the period left for validation. In conclusion, a forecast for the period of January to May 2015 and the adoption of the proposed model was able to provide a savings of 45% in relation to the scheme currently used by this HEI.
Nos últimos anos, diversos trabalhos foram publicados em relação à estimativa de variáveis relacionadas ao uso da energia elétrica, onde as mais variadas metodologias são utilizadas para realizar a modelagem e estimação da demanda por energia de países, Estados, empresas em geral e dos sistemas de ensino. Nesta dissertação foi escolhida esta última categoria e o foco consiste nas Instituições de Ensino Superior (IES). Procurando elaborar uma estimativa de Demanda Máxima de Ponta (DMP), mensal de potência de energia elétrica adequada às IESS, a partir da quantidade de alunos, e, se necessário, a partir de outras variáveis causais, que possam contribuir de maneira gerencial para a renegociação de contratos com concessionárias que levem à redução de custos anuais e que ainda podem contribuir para um melhor controle dos níveis de demanda máxima de energia elétrica. Para alcançar tal objetivo, foi realizada uma revisão da literatura a respeito de variáveis que poderiam apresentar correlação com a variável dependente DMP. Esta revisão indicou várias metodologias que poderiam contribuir para a solução do problema proposto: a Cadeia de Markov, a Metodologia de Regressão do vetor de Suporte, o Modelo de Programação Genética e as Redes Neurais Artificiais. Por ser uma metodologia menos complexa e direcionada a empresas de grande porte, adotou-se a Metodologia de Regressão Linear Múltipla (RLM). Foi selecionada uma IES e foram realizadas entrevistas com alguns engenheiros e técnico da sua divisão de engenharia elétrica, procurando entender melhor o uso da energia e o comportamento da variável DMP nesta IES, sendo disponibilizados os relatórios de energia do sistema de monitoração de energia onde os dados de DMP de janeiro de 2008 a dezembro de 2014 estavam contidos. Então, com base nestes dados e em pesquisa documental das candidatas a variáveis independentes, e, através da Metodologia (RLM), foi desenvolvido um modelo a partir dos dados de 72 meses, que teve seus resíduos avaliados, apresentando um coeficiente de determinação 𝑅2 igual a 0,883 .As variáveis independentes que permaneceram no modelo, a partir da utilização do método backward, foram 4(quatro) variáveis Dummy associadas a anos, seis variáveis deste tipo associadas a meses e uma variável fruto do produto entre dias letivos de graduação e quantidade de alunos da graduação matriculados. O modelo foi capaz de identificar a sazonalidade presente no comportamento da DMP da IES em estudo. Ele possibilitaria a contratação de DMP por mês, o que daria uma economia de 57% em relação ao modo de contratação tradicional (DMP fixo para todo o período), considerando o período de julho a dezembro, antes do período deixado para validação. Concluindo, foi realizada uma previsão para o período de janeiro a maio de 2015 e a adoção do modelo proposto foi capaz de proporcionar uma economia de 45% em relação ao esquema utilizado atualmente pela IES.

A república de Moçambique é um país localizado ao longo da costa Leste da África Austral, com a economia baseada essencialmente na prática da agricultura. A cultura do milho (Zea mays L.) é a mais importante, cultivada em regime de sequeiro, com rendimentos dependentes das condições meteorológicas. Modelos agrometeorológicos de estimativa de rendimentos de culturas alimentares são alternativas viáveis para tomada de decisão em medidas de segurança alimentar e abastecimento. O calendário agrícola e o sistema de produção tornam o uso de geotecnologias uma importante ferramenta para o monitoramento de culturas e o desenvolvimento de modelos de estimativa de rendimentos. Produtos de dados de sensoriamento remoto, como índices espectrais combinados com parâmetros agrometeorológicos podem melhorar as representações espaciais de rendimentos do milho em Moçambique. O ajuste de um modelo agrometeorológico espectral para estimativa de rendimentos do milho por regressão linear múltipla na província de Manica-Moçambique constituiu o objetivo do estudo. Foi realizado um mapeamento de áreas agrícolas por análise multitemporal do NDVI/MODIS e também foi avaliada a eficiência de variáveis agrometeorológicas e espectrais na estimativa de rendimentos do milho em uma área da província de Manica que envolve os distritos de Gondola, Manica, Mossurize e Sussundenga, responsáveis por mais de 80% da produção de milho na província nos anos de 2000 a 2009. Foi desenvolvido um modelo de início do ciclo do milho baseado em critérios de chuva, e estabelecendo um ciclo fixo do milho em 130 dias. A metodologia de mapeamento de áreas agrícolas consistiu em somatórios de imagens binárias geradas por diferença de NDVI máximo e mínimo ao longo do ciclo e estabelecimento de níveis de restrição com base em comparações com estatísticas oficiais por distrito. As variáveis agrometeorológicas testadas foram evapotranspiração relativa (ETr/ETm) e o índice de satisfação das necessidade de água (ISNA) calculados a partir de dados de estimativas de elementos meteorológicos do modelo do ECMWF. O conjunto de variáveis espectrais compreendiam composições de 16 dias de índices de vegetação EVI e NDVI provenientes do produto MOD13Q1 do sensor MODIS e o LSWI, gerado por diferença normalizada de bandas de refletância de superfície do infravermelho próximo e médio contidas no mesmo produto. O modelo agrometeorológico espectral envolveu as variáveis meteorológicas e espectrais como independentes sendo o rendimento médio e relativo, as variáveis dependentes ajustadas em um modelo de regressão múltipla. Todos os distritos, a exceção de Mossurize, geraram modelos com bom desempenho nas estimativas de rendimentos do milho e significado físico. O modelo regional, incluindo Gondola, Manica e Sussundenga e envolvendo o rendimento relativo foi o mais recomendado para estimativa de rendimentos do milho na região com r2 = 0,762 e RMSE de 9,46%.
Mozambique is a country located along the east coast of southern Africa, with an economy based primarily on agriculture. The Maize crop (Zea mays L.) is the most important crop, growing in rainfed conditions, with its yield dependent only on weather conditions. Agrometeorological models to forecast yields of food crops are viable alternatives for decision making on food safety measures and supply. The agricultural calendar and the production system make use of geotechnologies an important tool for crop monitoring and yield forecasting. Products from remote sensing data, combined with spectral indices and agrometeorological parameters can improve the spatial representations of maize yields in Mozambique. Setting an agrometeorological model to estimate the spectral yield of corn by multiple linear regression in Manica province, Mozambique was the objective of the study. Were conducted a mapping of agricultural areas by analyzing multitemporal NDVI / MODIS and also evaluated the effectiveness of spectral and meteorological variables in the estimated maize yield in an area of Manica province involving the districts of Gondola, Manica, Mossurize and Sussundenga responsible for more than 80% of corn production in the province in the years 2000 to 2009. A model was developed to estimate the beginnig of the corn cycle, using as a criteria the rainfall, and setting a fixed cycle of corn in 130 days. The methodology for mapping agricultural areas consisted of sums of binary images generated by the difference of maximum and minimum NDVI throughout the cycle and establishing levels of restriction based on comparisons with official statistics by district. Were tested the meteorological variables: the relative evapotranspiration (ETr / ETm) and the index of satisfaction of water needs (ISNA) calculated from data from meteorological model of ECMWF. The set of spectral variable were comprised of 16 days composition of vegetation indices NDVI and EVI from the MODIS product MOD13Q1 and LSWI generated from normalized difference of surface reflectance bands of near-infrared and medium infrared contained the same product. The meteorological and spectral variables was the set of independent variables and the average and relative yield were the set of dependent variables used to adjusted a multiple regression model, called agrometeorological-spectral model. To all districts, except for Mossurize were generated models with good performance in estimating the corn yield and with physical meaning. The regional model, including Gondola, Manica and Sussundenga and involving the relative yield was the most suitable for estimating corn yield in the region with r2 = 0.762 and RMSE of 9.46%.

There are two main parts in this thesis. The first one concerns valvometry, which is here the study of the distance between both parts of the shell of an oyster, over time. The health status of oysters can be characterized using valvometry in order to obtain insights about the quality of their environment. We consider that a renewal process with four states underlies the behaviour of the studied oysters. Such a hidden process can be retrieved from a valvometric signal by assuming that some probability density function linked with this signal, is bimodal. We then compare several estimators which take this assumption into account, including kernel density estimators.In another chapter, we compare several regression approaches, aiming at analysing transcriptomic data. To understand which explanatory variables have an effect on gene expressions, we apply a multiple testing procedure on these data, through the linear model FAMT. The SIR method may find nonlinear relations in such a context. It is however more commonly used when the response variable is univariate. A multivariate version of SIR was then developed. Procedures to measure gene expressions can be expensive. The sample size n of the corresponding datasets is then often small. That is why we also studied SIR when n is less than the number of explanatory variables p.

Submitted by JOSIANE SANTOS DE OLIVEIRA (josianeso) on 2017-05-30T11:58:58Z No. of bitstreams: 1 Luciano Lopes Vargas Amor_.pdf: 2139833 bytes, checksum: f496319c4cf4ede22ca6b861b4855664 (MD5)
Made available in DSpace on 2017-05-30T11:58:58Z (GMT). No. of bitstreams: 1 Luciano Lopes Vargas Amor_.pdf: 2139833 bytes, checksum: f496319c4cf4ede22ca6b861b4855664 (MD5) Previous issue date: 2017-01-12
CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
CNPQ – Conselho Nacional de Desenvolvimento Científico e Tecnológico
O setor da construção civil é hoje um grande gerador de resíduos, o que impacta negativamente o meio ambiente. Tanto a fase de projeto quanto a de execução de obras podem ocasionar essa geração. Quantificar o resíduo gerado em uma obra é fundamental para ter um controle do problema. A madeira, principal material empregado no uso provisório nos canteiros de obras, pode gerar grande volume de resíduo, sendo que o mesmo é descartado tão logo a obra termina, havendo pouco reaproveitamento ou reciclagem da mesma. Quantificar esse resíduo pode facilitar a gestão e o controle do mesmo, possibilitando a sustentabilidade de um material nobre e importante para o desenvolvimento da construção civil. O objetivo deste estudo consiste em propor um modelo para estimar a geração de resíduos de madeira de uso provisório em obras de edifícios verticais, analisando suas características de projeto e de execução. A estratégia de pesquisa utilizada foi o estudo de casos múltiplos, com coleta de dados de obras localizadas na região metropolitana de Porto Alegre, feitos através de entrevistas, observações in loco e análise de projetos. Ao todo, a amostra estudada foi formada por 22 obras de 11 empresas. Como resultado, foram gerados modelos estatísticos, a partir de regressão múltipla, contendo a variável dependente - volume de resíduos de madeira – e as variáveis independentes relativas ao projeto – número de pavimentos, número de pavimentos em subsolo, volume de concreto executado – e à execução - comprimento linear de tapume em madeira, índice de uso da madeira na confecção dos equipamentos de proteção coletiva e instalações provisórias. O modelo proposto possui R² ajustado de 0,893, indicando que é capaz de explicar 89,3% do fenômeno. As variáveis que mais impactaram foram o número de pavimentos e o número de pavimentos em subsolo. O modelo permite que as empresas construtoras possuam maior controle sobre o volume de resíduo de madeira gerado pelas suas obras, e possibilita realizar ações efetivas em prol da gestão dos resíduos, havendo maior controle e a redução dos mesmos. Possibilita também aplicá-lo em outras obras de diferentes regiões, considerando as características de projeto e de execução locais.
The construction industry is today a great generator of waste, which has a negative impact on the environment. Both the design and execution phases of works can cause this generation. To quantify the waste generated in a work is fundamental to control the problem. Wood, the main material used temporarily in construction sites, can generate a large volume of waste, which is discarded as soon as the work is finished, with little reuse or recycling. Quantifying this waste can facilitate the management and control of the same, allowing the sustainability of a material which is important and noble for the development of civil construction. The aim of this study is to propose a model to estimate the generation of wood waste used temporarily in vertical construction, analyzing their design and execution characteristics. The research strategy used was the study of multiple cases, with data collection of works located in the metropolitan region of Porto Alegre, made through interviews, in situ observations and project analysis In all, the sample studied was formed by 22 works from 11 companies. As a result, statistical models were generated from multiple regression, containing the dependent variable - volume of wood waste - and the independent variables related to the project - number of floors, number of floors in subsoil, volume of concrete executed - and execution - linear length of siding in wood, the index use of wood in the production of collective protection equipment and temporary facilities. The proposed model has an adjusted R² of 0.893, indicating that it is able to explain 89.3% of the phenomenon. The variables that most impacted it were the number of floors and the number of floors in subsoil. The model allows construction companies to have greater control over the volume of wood waste generated by their works, and enables effective actions to be taken in favor of waste management, with greater control and reduction of waste. It can also be applied in other works from different regions, considering the local design and execution characteristics.

Submitted by Nara Lays Domingues Viana Oliveira (naradv) on 2015-07-01T19:30:05Z No. of bitstreams: 1 micheledias.pdf: 1762771 bytes, checksum: ef597c67a9fa0e4be515580698ec6a19 (MD5)
Made available in DSpace on 2015-07-01T19:30:05Z (GMT). No. of bitstreams: 1 micheledias.pdf: 1762771 bytes, checksum: ef597c67a9fa0e4be515580698ec6a19 (MD5) Previous issue date: 2013-09-05
CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
FAPERGS - Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul
SINDUSCON/NH - Sindicato da Indústria da Construção Civil de Novo Hamburgo/RS
A significativa quantidade de resíduos gerada pelo setor da construção civil é hoje um grande problema ambiental a ser enfrentado pelo poder público e empresas construtoras. Para evitar ou reduzir a geração de resíduos é preciso atuar na origem, que pode estar desde o projeto até a fase de execução. O projeto consiste no passo inicial do empreendimento e tem influência significativa sobre o processo construtivo e produto final. Estudos apontam também que o sistema de gestão empregado por construtoras tem relação direta com a ocorrência de perdas, incluindo os resíduos. O conhecimento do índice de geração de resíduos é uma importante informação gerencial, com diferentes objetivos, desde a conscientização do problema, até para servir de base para o planejamento e implementação de ações voltadas à minimização da geração. Contudo, a quantificação dos resíduos de construção ainda pode ser considerada como um desafio, tendo em vista características peculiares da construção e a heterogeneidade dos resíduos. Observa-se muita variabilidade nos métodos empregados e variáveis consideradas, além das diferenças relativas à época e contextos pesquisados para geração dos índices e indicadores. Este trabalho propõe o desenvolvimento de um modelo estatístico para estimar a quantificação da geração de RCD em obras residenciais verticais, considerando características específicas das construtoras que atuam na região metropolitana de Porto Alegre-RS, investigando a influência do projeto e do sistema de produção. A estratégia de pesquisa consistiu num estudo de caso com múltiplas fontes de evidências, utilizando como ferramenta de análise estatística a regressão linear múltipla. A amostra é formada por dados de 20 obras residenciais verticais de 10 empresas que atuam na região metropolitana de Porto Alegre. Como resultado foi gerado um modelo estatístico, formado pela variável dependente (estimativa da quantidade de resíduo a ser gerada) e variáveis independentes relativas ao projeto (área do pavimento tipo, relação entre o número de pavimentos tipos e o número total de pavimentos e Índice econômico de compacidade) e ao sistema de produção (sistema construtivo e reaproveitamento de resíduos no canteiro). O modelo proposto obteve valor de R² ajustado = 0,69, permitindo explicar 69% da geração de resíduos em obras com características semelhantes, foi observada melhor aproximação da estimativa realizada pelo modelo quando comparado a algumas estimativas da literatura. A informação gerada através do modelo é importante na conscientização dos empresários locais e útil para basear ações no sentido de minimizar a geração, reciclar ou ações relativas ao destino final adequado, além de possibilitar comparações com informações sobre a geração de resíduos de outras localidades.
A significant amount of waste generated by the construction industry is now a major environmental problem to be faced by the government and construction companies. To avoid or reduce the generation of waste is necessary to work at the source, which may be from design to the implementation phase. The project consists in the initial step of the development and has significant influence on the construction process and final product. Studies also indicate that the management system employed by construction companies is directly related to the occurrence of losses, including waste. The knowledge of the waste generation rate is important management information, with different goals, since the awareness of the problem, even to serve as a basis for planning and implementation of actions aimed at minimizing the generation. However, the quantification of construction waste can still be considered as a challenge in view of the peculiar characteristics of the construction and the heterogeneity of the wastes. It is observed high variability in the methods employed and the variables considered, and the differences related to the time and contexts searched for generating indices and indicators. This work proposes the development of a statistical model to estimate the quantification of CDW generation in residential construction vertical, considering the specific characteristics of construction companies operating in the metropolitan area of Porto Alegre - RS, investigating the influence of design and production system. The research strategy consisted of a case study with multiple sources of evidence, using as a tool for statistical analysis with linear regression. The sample consists of data from 20 residential construction vertical of 10 companies operating in the metropolitan area of Porto Alegre. As a result generated a statistical model , formed by the dependent variable ( the estimated amount of waste to be generated ) and independent variables related to the project (area of the pavement type , relationship between the number of flooring types and the total number of floors and economic index compactness ) and production system ( building system and reuse of waste in construction ) . The proposed model obtained adjusted R² = 0.69 , allowing explain 69 % of waste generation in the works with similar characteristics was observed better approximation to the estimation made by the model when compared to some estimates of the literature . The information generated by the model is important in raising awareness of local entrepreneurs and useful to base actions to minimize the generation, recycling or actions relating to the appropriate destination, and enable comparisons with information on waste generation from other locations.

Submitted by Fabricia Fialho Reginato (fabriciar) on 2015-08-03T23:08:01Z No. of bitstreams: 1 JoiceCagliari.pdf: 1412838 bytes, checksum: 244299757ce9e71c1219d2018888f982 (MD5)
Made available in DSpace on 2015-08-03T23:08:01Z (GMT). No. of bitstreams: 1 JoiceCagliari.pdf: 1412838 bytes, checksum: 244299757ce9e71c1219d2018888f982 (MD5) Previous issue date: 2010
CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
O fósforo remanescente consiste na concentração de fósforo que permanece em solução após a agitação por 1 hora de uma amostra de solo com solução de CaCl2 0,01 mol L-1 contendo 60 mg L-1 P. O valor de fósforo remanescente pode ser utilizado como um bom indicador da capacidade de sorção aniônica de um solo haja vista ser mais sensível a sua composição mineralógica do que ao teor de sua fração argila. No Brasil, a utilização agronômica do fósforo remanescente é contemplada no sistema oficial de recomendação de fertilizantes e corretivos do Estado de Minas Gerais. O objetivo principal deste estudo foi o de desenvolver uma função de pedotransferência que permitisse estimar com razoável exatidão o valor de fósforo remanescente de solos representativos do Estado de São Paulo, a partir de outros atributos químicos de mais simples e/ou rotineira determinação laboratorial. Nesse contexto, duas funções de pedotransferência foram desenvolvidas com base em redes neurais artificiais (RNA) e análises de regressão linear múltipla (RLM), utilizando um banco de dados constituído por propriedades físicas e químicas de solos amostrados em diferentes localidades do Estado de São Paulo. As redes alimentadas adiante com múltiplas camadas foram utilizadas para desenvolver a função de pedotransferência baseada em RNA e a topologia foi determinada a partir de experimentos sucessivos. Os critérios de escolha da melhor rede neural foram, simultaneamente, o desempenho na etapa de treinamento, medido por meio do erro quadrático médio, e a capacidade de generalização, avaliada por meio de análises estatísticas entre os valores de Prem estimados e determinados analiticamente. A topologia da rede que melhor estimou o fósforo remanescente foi [3 14 1], ou seja, três neurônios na camada de entrada, quatorze em uma única camada intermediária e um na camada de saída; a função de ativação utilizada foi a sigmoidal logística, os valores de entrada foram normalizados entre [0;1] e o algoritmo de aprendizagem utilizado foi o resilient backpropagation. As três variáveis da camada de entrada foram o valor de pH medido em solução de NaF 1 mol L-1 (pH NaF), a soma de bases trocáveis (SB) e o teor de alumínio trocável (Al3+), sendo as duas últimas determinadas rotineiramente em análises de solo e a primeira de mais fácil e rápida obtenção que o fósforo remanescente. A função de pedotransferência baseada em RLM foi desenvolvida considerando as mesmas variáveis de entrada utilizadas na função de pedotransferência baseada em RNA. A comparação entre os desempenhos obtidos, para um mesmo conjunto de validação, mostrou que as funções de pedotransferência baseadas em redes neurais apresentam estimativas mais exatas do fósforo remanescente. Apesar do conjunto de dados utilizado não ser suficientemente abrangente para o estabelecimento de uma função de pedotransferência definitiva para a estimativa do fósforo remanescente, os resultados do presente trabalho indicam como promissor o desenvolvimento de um massivo banco de dados por meio do aproveitamento dos resultados analíticos continuamente gerados pelos vários laboratórios brasileiros dedicados à avaliação da fertilidade do solo e que contemple os valores de fósforo remanescente e pH NaF. Tal banco de dados permitirá o desenvolvimento de uma função de pedotransferência baseada em redes neurais artificiais cuja utilização possibilitará o cálculo imediato de valores suficientemente exatos de fósforo remanescente com razoável economia de recursos financeiros que seriam empregados na análise de um grande número de amostras.
The remaining phosphorus consists of the P concentration that remains in solution after shaking for 1 hour a soil sample with 0.01 mol L-1 CaCl2 containing 60 mg L-1 P. The remaining phosphorus values can be used as suitable indicators of the soil capacity of anion sorption due to be more dependable on the soil mineralogy than on the soil clay content. In Brazil, the remaining phosphorus is used as an ancillary variable in the official guidelines for determining fertilizer and amender requirements of agricultural soils of the Minas Gerais state. The main goal of this research was to develop a pedotransfer function (PTF) capable of providing fairly accurate estimates of remaining phosphorus values of representative soils of the São Paulo state from often-determined soil chemical properties and/or from other ones of easier determination. In this context, two pedotransfer functions were developed by using artificial neural networks (ANN) and multiple regression analysis (MRA) applied to a database formed by values of soil chemical and physical properties derived from soil surveys previously carried out in different locations of the São Paulo state. The multi-layer feedforward neural networks approach was used for the development of the ANN-based PTF being its topology determined from successive experiments. The simultaneous criteria adopted for choosing the best neural network were the performance during the training stage, measured by the mean squared error, and its capacity of providing accurate Prem values, which was evaluated by using a validation database in which statistical comparisons were done between the measured and estimated Prem values. The topology of the network that provided the most accurate estimates of the remaining phosphorus was [3 14 1], i.e., three neurons at the input layer, fourteen at a unique hidden layer and one neuron at the output layer; further development features included the use of the sigmoid logistic model as activation function, the input of data normalized in the [0;1] interval and the use of the resilient backpropagation learning algorithm. The three variables at the input layer were the soil pH value measured in 1 mol L-1 NaF (pH NaF), the sum of exchangeable bases (SB) and the soil content of exchangeable aluminum (Al3+), being the two last ones usually determined in soil test laboratories whereas the pH NaF determination is easier than the remaining phosphorus one. The MRA-based PTF was developed considering the same input variables of the ANN-based one, i.e., pH NaF, SB and Al3+. The comparisons performed with a same validation database showed that the pedotransfer function developed from artificial neural networks provided more accurate estimates of remaining phosphorus values. Despite the database used for the PTF development not be so comprehensive for the establishment of a definitive pedotransfer function for estimating remaining phosphorus values, the results of the present research indicate as promising the development of a massive database from chemical results often obtained by the Brazilian laboratories dedicated to soil fertility evaluation and that includes Prem and pH NaF values. This database will allow the development of a comprehensive ANN-based pedotransfer function capable of not only calculating suitable Prem values for practical applications but also reducing the expenses related to the analyses of a great number of soil samples.

Dans cette thèse, nous étudions le problème de réduction de la dimension dans le cadre du modèle de régression suivant Y=g(B X,e), où X est un vecteur de dimension p, Y appartient à R, la fonction g est inconnue et le bruit e est indépendant de X. Nous nous intéressons à l'estimation de la matrice B, de taille dxp où d est plus petit que p, (dont la connaissance permet d'obtenir de bonnes vitesses de convergence pour l'estimation de g). Ce problème est traité en utilisant deux approches distinctes. La première, appelée régression inverse nécessite la condition de linéarité sur X. La seconde, appelée semi-paramétrique ne requiert pas une telle condition mais seulement que X possède une densité lisse. Dans le cadre de la régression inverse, nous étudions deux familles de méthodes respectivement basées sur E[X f(Y)] et E[XX^T f(Y)]. Pour chacune de ces familles, nous obtenons les conditions sur f permettant une estimation exhaustive de B, aussi nous calculons la fonction f optimale par minimisation de la variance asymptotique. Dans le cadre de l'approche semi-paramétrique, nous proposons une méthode permettant l'estimation du gradient de la fonction de régression. Sous des hypothèses semi-paramétriques classiques, nous montrons la normalité asymptotique de notre estimateur et l'exhaustivité de l'estimation de B. Quel que soit l'approche considérée, une question fondamentale est soulevée : comment choisir la dimension de B ? Pour cela, nous proposons une méthode d'estimation du rang d'une matrice par test d'hypothèse bootstrap.

碩士
國立中興大學
土木工程學系
85
Multiple regression method provides a somewhat easy way of perdicting thedependent variable from knowledge of many other indendent variables and thus it can be applied to peak flow rate estimation in hydrology studies. Usually, the physiographic factors of a certain watershed respond dynamic features of stream characteristics and they also play important roles in determining runoffhydrograph at watershed outlet. This research thus uses geograph information system softwares ARC/INFO and ArcView to obtain watershed physiographic factorsand then uses multiple regression analysis to find the design runoff peak flowrate for different return periods. The results show that the design peak rate is highly related with the selected physiographic factors. Also, the regression result also reflects fairly good "hydrologic homogeneity" of the watershed innorthern and eastern part of Taiwan island. IN addition, the general framework of this study is constructed on ArcView GIS software. Because ArcView softwareemploys AVENUE object-oriented programming, the proposed methed also containsa group of user interfaces to automatically calculate the outcome regression calculation for different return periods which provide a user- friendly tool forthe users.

碩士
國立高雄應用科技大學
高階經營管理研究所碩士在職專班
99
Line loss rate of Taiwan Power Company（Taipower）is determined by dividing the difference between energy production & purchased and total energy sales by energy production & purchased, and it can be divided into three segments, primary transmission system loss rate, secondary transmission system loss rate and distribution system loss rate. The first step in the procedure for determining Taipower’s current monthly total system loss rate is to calculate both distribution system and secondary transmission loss, and then to obtain primary transmission loss rate sequentially. Due to the fact that meter reading factors can affect the accumulated energy sales (i.e. bimonthly meter reading for regular service customers and batch meter reading for low voltage customers), the monthly loss rate in the distribution system would vary dramatically, thus causing fluctuations of the monthly loss rate in the primary transmission accordingly and even obtaining irrational negative value. In this thesis, we collect the historical data of real power periodically for the whole year of 2010 for simulation via the new energy management system (EMS) which was implemented by Taipower in 2009. Multiple linear regression analysis, which selects energy production and purchased, power flow from central to north area, and power flow from south to central area as independent variables, and primary transmission system loss as dependent variable, is applied to derive Taipower’s monthly primary transmission system loss rate models. Based on these derived models, the monthly primary transmission system loss rate can be estimated accordingly and in combination with obtained secondary transmission system loss rate to calculate the distribution system loss rate finally. The proposed derived models can not only be brought on to improve the procedures for estimating total system loss and avoid irrational negative monthly primary transmission system loss rate but also predict future primary transmission system loss rate that can be taken as reference for determining the rational primary transmission system loss rate.

Regression models in medical research are widely used for prediction. When predicting the response at a future randomly chosen covariate vector x, problems can arise. The fit of a regression model to new data is nearly always worse than its fit to the original data, a deterioration called shrinkage. The Stein-type predictors give a uniformly lower expected mean squared error for prediction (MSEP) than least squares estimators under certain assumptions Different forms of the Stein-type and nonparametric shrinkage predictors were computed and compared using re-sampling and sample re-using. Data were generated from models with normally and nonnormally distributed error terms. Normally distributed data with residual variances following particular patterns of heteroscedasticity were investigated as well. The number of predictors, sample size, beta and the correlation between covariates were varied. Both the mean and median of the mean squared error for prediction (MSEP) and predicted residual sum of squares (PRESS) were computed from 10,000 simulations (5,000 for PRESS) for each shrinkage predictor and the results compared The results showed that the shrinkage predictors gave a uniformly lower MSEP/PRESS than the least squares predictors. The percentage of loss saving went up to 35% under certain conditions. The Stein-type predictors usually gave a lower MSEP/PRESS than the nonparametric form. One of the modified Stein-form/class estimators. This conclusion held even when moderate correlation existed in the covariates. The positive part estimator also dominated under different error structures The percentage of loss saving decreased as the sample size increased or the number of covariates decreased. The percentages of loss saving were fairly comparable when the covariates were either independent or had low (pairwise r = 0.1) correlation. When there were moderate correlations among covariates, the loss saving decreased significantly. Such findings were the same whether means or medians of MSEP and PRESS were considered
acase@tulane.edu