Predictive models for chronic renal disease using decision trees, naïve bayes and case-based methods

Data mining can be used in healthcare industry to “mine” clinical data to discover hidden information for intelligent and affective decision making. Discovery of hidden patterns and relationships often goes intact, yet advanced data mining techniques can be helpful as remedy to this scenario. This thesis mainly deals with Intelligent Prediction of Chronic Renal Disease (IPCRD). Data covers blood, urine test, and external symptoms applied to predict chronic renal disease. Data from the database is initially transformed to Weka (3.6) and Chi-Square method is used for features section. After normalizing data, three classifiers were applied and efficiency of output is evaluated. Mainly, three classifiers are analyzed: Decision Tree, Naïve Bayes, K-Nearest Neighbour algorithm. Results show that each technique has its unique strength in realizing the objectives of the defined mining goals. Efficiency of Decision Tree and KNN was almost same but Naïve Bayes proved a comparative edge over others. Further sensitivity and specificity tests are used as statistical measures to examine the performance of a binary classification. Sensitivity (also called recall rate in some fields) measures the proportion of actual positives which are correctly identified while Specificity measures the proportion of negatives which are correctly identified. CRISP-DM methodology is applied to build the mining models. It consists of six major phases: business understanding, data understanding, data preparation, modeling, evaluation, and deployment.

Parameter estimation of Poisson generalized linear mixed models based on three different statistical principles : a simulation study

Generalized linear mixed models are flexible tools for modeling non-normal data and are useful for accommodating overdispersion in Poisson regression models with random effects. Their main difficulty resides in the parameter estimation because there is no analytic solution for the maximization of the marginal likelihood. Many methods have been proposed for this purpose and many of them are implemented in software packages. The purpose of this study is to compare the performance of three different statistical principles - marginal likelihood, extended likelihood, Bayesian analysis-via simulation studies. Real data on contact wrestling are used for illustration.

Radiation properties of coil-coated steel in building envelope surfaces and the influence on building thermal performance

Recent studies have shown that the optical properties of building exterior surfaces are important in terms of energy use and thermal comfort. While the majority of the studies are related to exterior surfaces, the radiation properties of interior surfaces are less thoroughly investigated. Development in the coil-coating industries has now made it possible to allocate different optical properties for both exterior and interior surfaces of steel-clad buildings. The aim of this thesis is to investigate the influence of surface radiation properties with the focus on the thermal emittance of the interior surfaces, the modeling approaches and their consequences in the context of the building energy performance and indoor thermal environment. The study consists of both numerical and experimental investigations. The experimental investigations include parallel field measurements on three similar test cabins with different interior and exterior surface radiation properties in Borlänge, Sweden, and two ice rink arenas with normal and low emissive ceiling in Luleå, Sweden. The numerical methods include comparative simulations by the use of dynamic heat flux models, Building Energy Simulation (BES), Computational Fluid Dynamics (CFD) and a coupled model for BES and CFD. Several parametric studies and thermal performance analyses were carried out in combination with the different numerical methods. The parallel field measurements on the test cabins include the air, surface and radiation temperatures and energy use during passive and active (heating and cooling) measurements. Both measurement and comparative simulation results indicate an improvement in the indoor thermal environment when the interior surfaces have low emittance. In the ice rink arenas, surface and radiation temperature measurements indicate a considerable reduction in the ceiling-to-ice radiation by the use of low emittance surfaces, in agreement with a ceiling-toice radiation model using schematic dynamic heat flux calculations. The measurements in the test cabins indicate that the use of low emittance surfaces can increase the vertical indoor air temperature gradients depending on the time of day and outdoor conditions. This is in agreement with the transient CFD simulations having the boundary condition assigned on the exterior surfaces. The sensitivity analyses have been performed under different outdoor conditions and surface thermal radiation properties. The spatially resolved simulations indicate an increase in the air and surface temperature gradients by the use of low emittance coatings. This can allow for lower air temperature at the occupied zone during the summer. The combined effect of interior and exterior reflective coatings in terms of energy use has been investigated by the use of building energy simulation for different climates and internal heat loads. The results indicate possible energy savings by the smart choice of optical properties on interior and exterior surfaces of the building. Overall, it is concluded that the interior reflective coatings can contribute to building energy savings and improvement of the indoor thermal environment. This can be numerically investigated by the choice of appropriate models with respect to the level of detail and computational load. This thesis includes comparative simulations at different levels of detail.

Parameterization Of Turbulence Models Using 3DVAR Data Assimilation

In this research the 3DVAR data assimilation scheme is implemented in the numerical model DIVAST in order to optimize the performance of the numerical model by selecting an appropriate turbulence scheme and tuning its parameters. Two turbulence closure schemes: the Prandtl mixing length model and the two-equation k-ε model were incorporated into DIVAST and examined with respect to their universality of application, complexity of solutions, computational efficiency and numerical stability. A square harbour with one symmetrical entrance subject to tide-induced flows was selected to investigate the structure of turbulent flows. The experimental part of the research was conducted in a tidal basin. A significant advantage of such laboratory experiment is a fully controlled environment where domain setup and forcing are user-defined. The research shows that the Prandtl mixing length model and the two-equation k-ε model, with default parameterization predefined according to literature recommendations, overestimate eddy viscosity which in turn results in a significant underestimation of velocity magnitudes in the harbour. The data assimilation of the model-predicted velocity and laboratory observations significantly improves model predictions for both turbulence models by adjusting modelled flows in the harbour to match de-errored observations. 3DVAR allows also to identify and quantify shortcomings of the numerical model. Such comprehensive analysis gives an optimal solution based on which numerical model parameters can be estimated. The process of turbulence model optimization by reparameterization and tuning towards optimal state led to new constants that may be potentially applied to complex turbulent flows, such as rapidly developing flows or recirculating flows.

Forecasting accuracy and estimation uncertainty using VAR models with short- and long-term economic restrictions: a Monte-Carlo study

Using vector autoregressive (VAR) models and Monte-Carlo simulation methods we investigate the potential gains for forecasting accuracy and estimation uncertainty of two commonly used restrictions arising from economic relationships. The Örst reduces parameter space by imposing long-term restrictions on the behavior of economic variables as discussed by the literature on cointegration, and the second reduces parameter space by imposing short-term restrictions as discussed by the literature on serial-correlation common features (SCCF). Our simulations cover three important issues on model building, estimation, and forecasting. First, we examine the performance of standard and modiÖed information criteria in choosing lag length for cointegrated VARs with SCCF restrictions. Second, we provide a comparison of forecasting accuracy of Ötted VARs when only cointegration restrictions are imposed and when cointegration and SCCF restrictions are jointly imposed. Third, we propose a new estimation algorithm where short- and long-term restrictions interact to estimate the cointegrating and the cofeature spaces respectively. We have three basic results. First, ignoring SCCF restrictions has a high cost in terms of model selection, because standard information criteria chooses too frequently inconsistent models, with too small a lag length. Criteria selecting lag and rank simultaneously have a superior performance in this case. Second, this translates into a superior forecasting performance of the restricted VECM over the VECM, with important improvements in forecasting accuracy ñreaching more than 100% in extreme cases. Third, the new algorithm proposed here fares very well in terms of parameter estimation, even when we consider the estimation of long-term parameters, opening up the discussion of joint estimation of short- and long-term parameters in VAR models.

Graphical models and point set matching

Point pattern matching in Euclidean Spaces is one of the fundamental problems in Pattern Recognition, having applications ranging from Computer Vision to Computational Chemistry. Whenever two complex patterns are encoded by two sets of points identifying their key features, their comparison can be seen as a point pattern matching problem. This work proposes a single approach to both exact and inexact point set matching in Euclidean Spaces of arbitrary dimension. In the case of exact matching, it is assured to find an optimal solution. For inexact matching (when noise is involved), experimental results confirm the validity of the approach. We start by regarding point pattern matching as a weighted graph matching problem. We then formulate the weighted graph matching problem as one of Bayesian inference in a probabilistic graphical model. By exploiting the existence of fundamental constraints in patterns embedded in Euclidean Spaces, we prove that for exact point set matching a simple graphical model is equivalent to the full model. It is possible to show that exact probabilistic inference in this simple model has polynomial time complexity with respect to the number of elements in the patterns to be matched. This gives rise to a technique that for exact matching provably finds a global optimum in polynomial time for any dimensionality of the underlying Euclidean Space. Computational experiments comparing this technique with well-known probabilistic relaxation labeling show significant performance improvement for inexact matching. The proposed approach is significantly more robust under augmentation of the sizes of the involved patterns. In the absence of noise, the results are always perfect.

Diversificação e performance - uma análise das estratégias de diversificação em empresas brasileiras

Este trabalho visa explorar, com base em dados brasileiros, a relação entre diversificação e performance. Como medida de performance serão utilizados valores correspondentes ao índice q de Tobin para empresas de capital aberto. Para o cálculo do índice de diversificação de uma firma serão utilizados índices compostos a partir da codificação americana SIC (Standard Industry Code). A verificação da relação estatística entre diversificação e performance será então aferida através da aplicação de modelos de regressão linear e sistemas de equações estruturais simultâneas.

Acumulação tecnológica e aprimoramento de performance competitiva: evidências de diferenças e similaridades e médias empresas do setor metal-mecânico no estado do rio de janeiro

The main objective of this dissertation is to examine the implications of technological capacities in the improvement of technical performance indexes, specifically at the company level. These relationships were examined in a small sample of metal-working enterprises in the state of Rio de Janeiro (1960 to 2006). Although diverse studies on technological competences have been carried out in the last twenty years, a gap in empirical studies still exist that correlate the performance of companies in the context of developing countries, especially in Brazil. Aiming to contribute to a reduction of these gaps, this dissertation examines the questions by the light of available models in literature, which opting themselves to using operational indexes of companies. For drawing the accumulation of technological competences in this study, the metric proposal by Figueiredo (2000) shall be used indicating the levels of technological qualification in process, product, and equipment functions. The empirical evidence examined in this dissertation is both qualitative and quantitative in nature and were collected, first hand, through extensive field research involving informal interviews, meetings, direct-site observation and document analysis. In relation to the results, the evidence suggests that: - In terms of technological accumulation, a company reached Level 5 of technological capacity in process and organization of production as well as product and equipment. Three companies obtained Level 4 in the function process function while two others had reached the same technological level in the functions of product and equipment. Two companies had reached Level 3 in the product and equipment functions and one remained this level in the function of process; - In terms of the rate of accumulation of technological capacities, the observed companies had reached Level 4 needs 29 years in process function, 32 years in product function and 29 years in equipment function; - In terms of improvement performance pointers, a company which reached Level 5 of technological capacity improved in 70% of its indicators of performance, while the company that had achieved Level 4 had raised its pointers 60% and the other companies had gotten improved in the order of 40%. It was evidenced that the majority of the pointers of the companies with higher levels of technological capacities had obtained better performance. This dissertation contributes to advancing the strategic management of companies in metal-working segment to understanding internal accumulation of technological capacity and indicators of performance especially in the field of empirical context studied. This information offers management examples of how to improve competitive performance through the accumulation of technological capacities in the process, product and equipment functions.

Evaluating Value-at-Risk models via Quantile regressions

This paper is concerned with evaluating value at risk estimates. It is well known that using only binary variables to do this sacrifices too much information. However, most of the specification tests (also called backtests) avaliable in the literature, such as Christoffersen (1998) and Engle and Maganelli (2004) are based on such variables. In this paper we propose a new backtest that does not realy solely on binary variable. It is show that the new backtest provides a sufficiant condition to assess the performance of a quantile model whereas the existing ones do not. The proposed methodology allows us to identify periods of an increased risk exposure based on a quantile regression model (Koenker & Xiao, 2002). Our theorical findings are corroborated through a monte Carlo simulation and an empirical exercise with daily S&P500 time series.

Model selection, estimation and forecasting in VAR models with short-run and long-run restrictions

We study the joint determination of the lag length, the dimension of the cointegrating space and the rank of the matrix of short-run parameters of a vector autoregressive (VAR) model using model selection criteria. We suggest a new two-step model selection procedure which is a hybrid of traditional criteria and criteria with data-dependant penalties and we prove its consistency. A Monte Carlo study explores the finite sample performance of this procedure and evaluates the forecasting accuracy of models selected by this procedure. Two empirical applications confirm the usefulness of the model selection procedure proposed here for forecasting.

Estimation of DSGE Models: A Monte Carlo Analysis

Neste trabalho investigamos as propriedades em pequena amostra e a robustez das estimativas dos parâmetros de modelos DSGE. Tomamos o modelo de Smets and Wouters (2007) como base e avaliamos a performance de dois procedimentos de estimação: Método dos Momentos Simulados (MMS) e Máxima Verossimilhança (MV). Examinamos a distribuição empírica das estimativas dos parâmetros e sua implicação para as análises de impulso-resposta e decomposição de variância nos casos de especificação correta e má especificação. Nossos resultados apontam para um desempenho ruim de MMS e alguns padrões de viés nas análises de impulso-resposta e decomposição de variância com estimativas de MV nos casos de má especificação considerados.

Persistência na performance de fundos de investimento imobiliário brasileiros entre 2008 e 2012

O objetivo deste trabalho é avaliar a persistência na performance de fundos de investimento imobiliário. Para isso, adotamos metodologia semelhante à de Carhart (1997): analisamos o desempenho, ao longo do tempo, de carteiras de fundos selecionados segundo seus percentis de retorno. Posteriormente, a performance desses fundos foi avaliada através de modelos multifatores. Para determinação desses modelos, foram criados índices baseados nas informações das ações do mercado brasileiro e do IBOVESPA. Os resultados sugerem que fundos de investimento imobiliário de retorno superior apresentam persistência em seus desempenhos. No caso dos modelos multifatores, conclui-se que os dados utilizados do mercado acionário não representam de forma satisfatória fundos de investimento imobiliário. No entanto, quando a modelagem é feita utilizando-se o IFIX, nota-se alfa positivo para os fundos mais rentáveis.

Consolidação bancária e a performance dos bancos pequenos no Brasil

A indústria bancária brasileira foi transformada nas últimas décadas em meio a um fenômeno conhecido como consolidação, que marca uma concentração do mercado em poucas instituições. O objetivo do trabalho é testar empiricamente quais as causas desse processo no Brasil. As duas hipóteses testadas foram formuladas por Berger, Dick et al. (2007): a hipótese da eficiência indica que avanços tecnológicos melhoram a competitividade dos grandes em relação aos pequenos. Deste modo, os resultados dos pequenos são sacrificados por esse fator. Por outro lado, a hipótese da arrogância afirma que os administradores realizam fusões e aquisições pelos maiores bônus dos grandes conglomerados, mas as deseconomias de escala são superiores aos ganhos competitivos da tecnologia e, com o tempo, os pequenos passam a competir em vantagem. Modelos de dados em painel foram utilizados para testar se houve pressões competitivas durante o processo de consolidação. A conclusão foi de que a hipótese da eficiência explica melhor empiricamente o fenômeno brasileiro, assim como o norte-americano. A pressão para diminuição de receitas financeiras foi o fator determinante para que os bancos pequenos sofressem efeitos deletérios com o aumento do peso dos grandes na indústria.

A moderação da concentração setorial no efeito do ambiente institucional na performance empresarial

O presente estudo avança a compreensão da performance empresarial ao propor que condições dos setores, especificamente a concentração setorial, moderam a relação entre instituições e desempenho das firmas. Já é sabido que o ambiente institucional impacta o desempenho das firmas (Makino, Isobe, & Chan, 2004) e que as reformas pró-mercado contribuem para que esse efeito seja positivo, tanto para firmas domésticas como estrangeiras (Cuervo-Cazurra & Dau, 2009). A explicação desse efeito é baseada na economia dos custos de transação (Coase, 1937; Commons, 1934). Contudo, não se sabe se o efeito é o mesmo para todos os setores e se fatores moderam a relação. Esta tese contou com 230.222 observações referentes a 10.903 empresas em 64 países em um intervalo de 23 anos coletados em diferentes bancos de dados. Foi testada a interação de seis variáveis institucionais com o índice Herfindahl-Hirschman (HHI) para três variáveis dependentes diferentes: retorno sobre ativos (ROA), retorno sobre patrimônio líquido (ROE) e crescimento de vendas composto de três anos. Duas estratégias empíricas foram utilizadas: modelos com efeitos fixos e modelos hierárquicos (multinível). Os resultados confirmaram a hipótese de que a interação do HHI é significante e negativa com quatro variáveis institucionais: voice and accountability, efetividade do governo, qualidade regulatória e controle da corrupção. Concentração setorial modera o efeito do ambiente institucional na performance empresarial. Em contextos onde as instituições são sólidas, a força de agentes como sindicatos, associações, imprensa e consumidor assume poder de barganha, refreando o poder das empresas e o oportunismo. Regras legais, direito comum e instituições tendem a limitar o poder unilateral em relações contratuais de todos os tipos, independe da fonte do poder (Macneil, 1980). Observou-se adicionalmente que a proteção ao oportunismo se dá principalmente por meio das instituições informais, como a proteção à democracia, direitos do consumidor e controle da corrupção. Ao propiciar poder aos outros agentes que interagem com as empresas, instituições fortes garantem o enforcement de compromissos contratuais, em particular os contratos sociais (Argyres & Liebeskind, 1999). Como implicações, essa tese propõe que estratégias de expansão dentro do setor, aquisição de participação de mercado e fusões e aquisições dentro do setor são mais adequadas em ambientes institucionais mais fracos que em ambientes fortes. Empresas que possuem alta participação de mercado devem reconhecer o impacto negativo que podem sofrer em seu desempenho a partir de melhorias institucionais. Finalmente, o estudo reforça a importância do reconhecimento por parte de governos de que setores e firmas se beneficiam de forma desigual das mudanças institucionais. O conhecimento prévio desses impactos pode servir de direcionamento para a formulação de políticas públicas justas e eficientes. As principais limitações estão relacionadas à base de dados, exclusivamente composta de empresas com capital aberto, a forma pela qual a classificação de algumas empresas diversificadas foi feita e o fato dessa tese não investigar diretamente o poder de barganha nem ao menos o oportunismo, mas somente o poder moderador da concentração setorial no efeito das instituições no desempenho.

Reality check for volatility models

Asset allocation decisions and value at risk calculations rely strongly on volatility estimates. Volatility measures such as rolling window, EWMA, GARCH and stochastic volatility are used in practice. GARCH and EWMA type models that incorporate the dynamic structure of volatility and are capable of forecasting future behavior of risk should perform better than constant, rolling window volatility models. For the same asset the model that is the ‘best’ according to some criterion can change from period to period. We use the reality check test∗ to verify if one model out-performs others over a class of re-sampled time-series data. The test is based on re-sampling the data using stationary bootstrapping. For each re-sample we check the ‘best’ model according to two criteria and analyze the distribution of the performance statistics. We compare constant volatility, EWMA and GARCH models using a quadratic utility function and a risk management measurement as comparison criteria. No model consistently out-performs the benchmark.