No description available.

M.Com. (Business Management) / Forecasting is an important function used in a wide range of business planning or decision-making situations. The purpose ofthis study was to build a sales forecasting model that would be practical and cost effective, from the various forecasting methods and techniques available. Various forecast models, methods and techniques are outlined in the initial part of this study by the author. The author has outlined some of the fundamentals and limitations that underline the preparations of forecasting models. It is not the purpose of this study to microscopically dissect each forecasting model, method or technique. Various forecasting options were assessed in a manner that could provide some relevance to the study, thus providing a general framework for the construction of the specific sales forecasting model. Appropriate data sources were described and analysed. The data was further tested using the author's chosen quantitative forecasting techniques. Results were interpreted, and included into the author's untested sales model. It is the author's opinion that the sales model is practical, cost effective and gives a general sales forecast.

The global climate is changing due to the accumulation of greenhouse gases (GHGs) in the atmosphere, primarily due to anthropogenic activity. The dominant GHG is CO₂ which originates from combustion of fossil fuels, land use change and management. The terrestrial biosphere is a key driver of climate and biogeochemical cycles at regional and global scales. Furthermore, the response of the Earth system to future drivers of climate change will depend on feedbacks between biogeochemistry and climate. Therefore, understanding these processes requires a mechanistic approach in any model simulation framework. However ecosystem processes are complex and nonlinear and consequently models need to be validated against observations at multiple spatial scales. In this thesis the weather research and forecasting model (WRF) has been coupled to the mechanistic terrestrial ecosystem model soil-plant-atmosphere (SPA), creating WRF-SPA. The thesis is split into three main chapters: i. WRF-SPA model development and validation at multiple spatial scales, scaling from surface fluxes of CO₂ and energy to aircraft profiles and tall tower observations of atmospheric CO₂ concentrations. ii. Investigation of ecosystem contributions to observations of atmospheric CO₂ concentrations made at tall tower Angus, Dundee, Scotland using ecosystem specific CO₂ tracers at seasonal and interannual time scales. iii. An assessment of detectability of a policy relevant national scale afforestation by observations made at a tall tower. Detectability of changes in atmospheric CO₂ concentrations was assessed through a comparison of a control simulation, using current day forest extent, and an experimentally afforested simulation using WRF-SPA. WRF-SPA performs well at both site and regional scales, accurately simulating aircraft profiles of CO₂ concentration magnitudes (error <+- 4 ppm), indicating appropriate source sink distribution and realistic atmospheric transport. Hourly observations made at tall tower Angus were also well simulated by WRF-SPA (R² = 0.67, RMSE = 3.5 ppm, bias = 0.58 ppm). Analysis of CO₂ tracers at tall tower Angus show an increase in the seasonal error between WRF-SPA simulated atmospheric CO₂ and observations, which coincides with simulated cropland harvest. WRF-SPA does not simulate uncultivated land associated with agriculture, which in Scotland represents 36 % of agricultural holdings. Therefore, uncultivated land components may provide an explanation for the increase in model-data error. Interannual variation in weather is indicated to have a greater impact on ecosystem specific contributions to atmospheric CO₂ concentrations at Angus than variation in surface activity. In a model experiment, afforestation of Scotland was simulated to test the impact on Scotland’s carbon balance. The changes were shown to be potentially detectable by observations made at tall tower Angus. Afforestation results in a reduction in atmospheric CO₂ concentrations by up to 0.6 ppm at seasonal time scales at tall tower Angus. Detection of changes in forest surface net CO₂ uptake flux due to afforestation was improved through the use of a network of tall towers (R² = 0.83) compared to tall tower Angus alone (R² = 0.75).

The launch of large stratospheric balloons is highly dependant on the meteorological conditions at ground level, including wind speed. The balloon launch base Esrange Space Center in northern Sweden currently uses forecasts delivered through the Swedish Meteorological and Hydrological Institute to predict opportunities for balloon launches. However the staff at Esrange Space Center experience that the current forecasts are not accurate enough. For that reason the Weather Research and Forecasting model is used to improve the forecast. The model performs a Large Eddy Simulation over the area closest to Esrange Space Center to predict wind speed and turbulence. During twelve hypothetical launch days the improved forecast have an overall accuracy of 93% compared to the old forecast accuracy of 69%. With some improvements and the right computational power the system is thought to be operationally viable.

Background: Dengue is a viral infectious disease that is transmitted by mosquitoes. The disease causes a significant health burden in tropical countries, and has been a public health burden in Singapore for several decades. Severe complications such as hemorrhage can develop and lead to fatal outcomes. Before tetravalent vaccine and drugs are available, vector control is the key component to control dengue transmission. Vector control activities need to be guided by surveillance of outbreak and implement timely action to suppress dengue transmission and limit the risk of further spread. This study aims to explore the feasibility of developing a dengue early warning system using temperature and rainfall as main predictors. The objectives were to 1) analyze the relationship between dengue cases and weather predictors, 2) identify the optimal lead time required for a dengue early warning, 3) develop forecasting models, and 4) translate forecasts to dengue risk indices. Methods: Poisson multivariate regression models were established to analyze relative risks of dengue corresponding to each unit change of weekly mean temperature and cumulative rainfall at lag of 1-20 weeks. Duration of vector control for localized outbreaks was analyzed to identify the time required by local authority to respond to an early warning. Then, dengue forecasting models were developed using Poisson multivariate regression. Autoregression, trend, and seasonality were considered in the models to account for risk factors other than temperature and rainfall. Model selection and validation were performed using various statistical methods. Forecast precision was analyzed using cross-validation, Receiver Operating Characteristics curve, and root mean square errors. Finally, forecasts were translated into stratified dengue risk indices in time series formats. Results: Findings showed weekly mean temperature and cumulative rainfall preceded higher relative risk of dengue by 9-16 weeks and that a forecast with at least 3 months would provide sufficient time for mitigation in Singapore. Results showed possibility of predicting dengue cases 1-16 weeks using temperature and rainfall; whereas, consideration of autoregression and trend further enhance forecast precision. Sensitivity analysis showed the forecasting models could detect outbreak and non-outbreak at above 90% with less than 20% false positive. Forecasts were translated into stratified dengue risk indices using color codes and indices ranging from 1-10 in calendar or time sequence formats. Simplified risk indices interpreted forecast according to annual alert and outbreak thresholds; thus, provided uniform interpretation. Significance: A prediction model was developed that forecasted a prognosis of dengue up to 16 weeks in advance with sufficient accuracy. Such a prognosis can be used as an early warning to enhance evidence-based decision making and effective use of public health resources as well as improved effectiveness of dengue surveillance and control. Simple and clear dengue risk indices improve communications to stakeholders.

Electricity price forecasting is an important task for electricity market participants since the very beginning of the deregulation. Accurate forecasting is essential for designing bidding strategy, risk management, and market operation. However, due to the compli-cated factors affecting electricity prices, there are more uncertainties in electricity price forecasting and hence more complex than demand forecasting. This makes accurate price forecasting very difficult. In the last decade, several methods have been developed in order to fully capture the peculiarities of electricity price dynamics, from classic econometric time series models, e.g., autoregressive moving average (ARMA) model, generalized autoregressive conditional heteroscedasticity (GARCH) model to modern machine learning based techniques such as artificial neural networks (ANN) and sup-port vector machine (SVM). In spite of all models proposed in the literature, there is still no clear consensus about which model is substantively outperforming others. Therefore, when a single method is used, decision-makers are facing the risk of not choosing the best one. On the other hand, the prediction of electricity market prices still involves large errors. If decision-makers take the prediction result on faith, prediction errors could exposure them to serious financial risks. Based on these findings, it can conclude that (1) systematic methodologies and implementations which can efficiently address model selection uncertainty in price forecasting require an investigation; (2) more powerful and robust price forecasting models are still needed to reduce the fore-cast errors; and (3) In addition, the emphasis of price forecasting should shift away from point forecast to uncertainty around the forecast. Unfortunately, most researches in this area have been devoted to finding the single “best” estimates rather than dealing with the uncertainty in model selection and quantifying the predictive uncertainty. In this thesis the research focus is on: (1) finding methodologies and efficient imple-mentations to deal with the uncertainty in model selection; (2) developing more power-ful machine learning based approaches to model electricity spot prices and further im-proving the accuracy of electricity market price forecast; and (3) incorporating uncer-tainty estimation into the application of price forecasting. The thesis makes three main contributions to the study of this topic. Firstly, it proposes linear, nonlinear forecast combination frameworks to deal with model selection prob-lem; secondly, it introduces two novel models: support vector machine based nonlinear generalized autoregressive conditional heteroscedasticity model (SVM-GARCH) and extreme learning machine (ELM) to the price forecasting and furthermore gives a series of bootstrap-based interval construction procedures to quantify the prediction uncer-tainty. Finally, it proposes a more robust interval forecasting approach which is based on quantile regression to electricity price forecasting literature. The effectiveness and efficiency of the proposed approaches have been tested based on real market data of Australian National Electricity Market (NEM).

Electricity price forecasting is an important task for electricity market participants since the very beginning of the deregulation. Accurate forecasting is essential for designing bidding strategy, risk management, and market operation. However, due to the compli-cated factors affecting electricity prices, there are more uncertainties in electricity price forecasting and hence more complex than demand forecasting. This makes accurate price forecasting very difficult. In the last decade, several methods have been developed in order to fully capture the peculiarities of electricity price dynamics, from classic econometric time series models, e.g., autoregressive moving average (ARMA) model, generalized autoregressive conditional heteroscedasticity (GARCH) model to modern machine learning based techniques such as artificial neural networks (ANN) and sup-port vector machine (SVM). In spite of all models proposed in the literature, there is still no clear consensus about which model is substantively outperforming others. Therefore, when a single method is used, decision-makers are facing the risk of not choosing the best one. On the other hand, the prediction of electricity market prices still involves large errors. If decision-makers take the prediction result on faith, prediction errors could exposure them to serious financial risks. Based on these findings, it can conclude that (1) systematic methodologies and implementations which can efficiently address model selection uncertainty in price forecasting require an investigation; (2) more powerful and robust price forecasting models are still needed to reduce the fore-cast errors; and (3) In addition, the emphasis of price forecasting should shift away from point forecast to uncertainty around the forecast. Unfortunately, most researches in this area have been devoted to finding the single “best” estimates rather than dealing with the uncertainty in model selection and quantifying the predictive uncertainty. In this thesis the research focus is on: (1) finding methodologies and efficient imple-mentations to deal with the uncertainty in model selection; (2) developing more power-ful machine learning based approaches to model electricity spot prices and further im-proving the accuracy of electricity market price forecast; and (3) incorporating uncer-tainty estimation into the application of price forecasting. The thesis makes three main contributions to the study of this topic. Firstly, it proposes linear, nonlinear forecast combination frameworks to deal with model selection prob-lem; secondly, it introduces two novel models: support vector machine based nonlinear generalized autoregressive conditional heteroscedasticity model (SVM-GARCH) and extreme learning machine (ELM) to the price forecasting and furthermore gives a series of bootstrap-based interval construction procedures to quantify the prediction uncer-tainty. Finally, it proposes a more robust interval forecasting approach which is based on quantile regression to electricity price forecasting literature. The effectiveness and efficiency of the proposed approaches have been tested based on real market data of Australian National Electricity Market (NEM).

A presente tese tem como objetivo propor um modelo de previsão para estimar o lucro médio esperado na concessão de crédito para pessoas físicas em empresas comerciais, obtendo assim uma medida monetária para dar suporte à tomada de decisão. O modelo proposto foi desenvolvido em três grandes etapas: 1) pré-processamento; 2) modelos de classificação; e 3) modelo de previsão do risco monetário. A primeira etapa inclui três passos: (i) delimitação da população, (ii) seleção da amostra, e (iii) análise preliminar. Na segunda etapa mais dois passos são necessários: (i) construção dos modelos, e (ii) qualidade dos modelos. Por fim, a última etapa trata das definições para construção do modelo de previsão do risco monetário propriamente dito, que utilizou os seguintes métodos: (i) ensemble, (ii) hybrid, e (iii) regressão linear múltipla. A exequibilidade do modelo proposto foi testada em dados reais de concessão de crédito. São avaliados os resultados de utilização do modelo de previsão, de forma a verificar o potencial aumento nos ganhos a partir da concessão do crédito, comparando quatro cenários: (i) sem utilizar nenhum modelo de previsão de risco de crédito; (ii) utilizando o modelo de classificação obtido com a regressão logística; (iii) utilizando o modelo de classificação obtido com a rede neural; e (iv) utilizando o modelo proposto para previsão do risco monetário. O modelo construído demonstrou resultados promissores na previsão do lucro médio esperado, apresentando um aumento estimado de 94,97% em comparação com o cenário sem uso de modelo de previsão, e um aumento de 26,08% quando comparado com o cenário de uso do modelo de classificação obtido com regressão logística. Uma análise de sensibilidade dos resultados com variações na margem de lucro por transação também foi realizada, evidenciando sua robustez. Nesse sentido, o modelo proposto se mostra efetivo como ferramenta de apoio para gestão no processo de decisão de concessão de crédito. / This thesis aims to propose a forecasting model to estimate the expected average profit in lending to individuals in commercial companies, thus obtaining a monetary measure to support decision making. The proposed model was developed in three major stages: 1) preprocessing, 2) classification models, and 3) model to forecast the currency risk. The first stage includes three steps: (i) delimitation of the population, (ii) sample selection, and (iii) preliminary analysis. In the second stage two more steps are necessary: (i) construction of models, and (ii) quality of the models. Finally, the last stage is regarding to the definitions for the construction of model prediction of the currency risk itself, which used the following methods: (i) ensemble, (ii) hybrid, and (iii) multiple linear regressions. The feasibility of the proposed model was tested on real data of grant credit. Results are evaluated using the prediction model in order to verify the potential increase in profits from the grant credit, comparing four scenarios: (i) without using any prevision model of credit risk, (ii) using the classification model obtained by logistic regression, (iii) using the classification model obtained with the neural network, and (iv) using the model to forecast the currency risk. The constructed model showed promising results in predicting the expected average profits, with an estimated increase of 94.97% compared to the scenario without the use of forecasting model, and an increase of 26.08% compared with the scenario of the classification model obtained by logistic regression. A sensitivity analysis of the results with variations in the profit margin per transaction was also performed, demonstrating its robustness. Accordingly, the proposed model proved effective as a support tool for management in the decision to grant credit.

abstract: This research work uses the Weather Research and Forecasting Model to study the effect of large wind farms with an area of 900 square kilometers and a high power density of 7.58 W/m2 on regional climate. Simulations were performed with a wind farm parameterization scheme turned on in south Oregon. Control cases were also run with the parameterization scheme turned off. The primary emphasis was on offshore wind farms. Some analysis on onshore wind farms was also performed. The effects of these wind farms were studied on the vertical profiles of temperature, wind speed, and moisture as well as on temperature and on wind speed near the surface and at hub height. The effects during the day and at night were compared. Seasonal variations were also studied by performing simulations in January and in July. It was seen that wind farms produce a reduction in wind speed at hub height and that the downward propagation of this reduction in wind speed lessens as the atmosphere becomes more stable. In all the cases studied, the wind farms produced a warming effect near the surface, with greater atmospheric stability leading to higher near-surface temperatures. It was also observed that wind farms caused a drying effect below the hub height and a moistening effect above it, because they had facilitated vertical transport of moisture in the air from the lower layers of the atmosphere to the layers of the atmosphere above the wind farm. / Dissertation/Thesis / Masters Thesis Mechanical Engineering 2016

A presente tese tem como objetivo propor um modelo de previsão para estimar o lucro médio esperado na concessão de crédito para pessoas físicas em empresas comerciais, obtendo assim uma medida monetária para dar suporte à tomada de decisão. O modelo proposto foi desenvolvido em três grandes etapas: 1) pré-processamento; 2) modelos de classificação; e 3) modelo de previsão do risco monetário. A primeira etapa inclui três passos: (i) delimitação da população, (ii) seleção da amostra, e (iii) análise preliminar. Na segunda etapa mais dois passos são necessários: (i) construção dos modelos, e (ii) qualidade dos modelos. Por fim, a última etapa trata das definições para construção do modelo de previsão do risco monetário propriamente dito, que utilizou os seguintes métodos: (i) ensemble, (ii) hybrid, e (iii) regressão linear múltipla. A exequibilidade do modelo proposto foi testada em dados reais de concessão de crédito. São avaliados os resultados de utilização do modelo de previsão, de forma a verificar o potencial aumento nos ganhos a partir da concessão do crédito, comparando quatro cenários: (i) sem utilizar nenhum modelo de previsão de risco de crédito; (ii) utilizando o modelo de classificação obtido com a regressão logística; (iii) utilizando o modelo de classificação obtido com a rede neural; e (iv) utilizando o modelo proposto para previsão do risco monetário. O modelo construído demonstrou resultados promissores na previsão do lucro médio esperado, apresentando um aumento estimado de 94,97% em comparação com o cenário sem uso de modelo de previsão, e um aumento de 26,08% quando comparado com o cenário de uso do modelo de classificação obtido com regressão logística. Uma análise de sensibilidade dos resultados com variações na margem de lucro por transação também foi realizada, evidenciando sua robustez. Nesse sentido, o modelo proposto se mostra efetivo como ferramenta de apoio para gestão no processo de decisão de concessão de crédito. / This thesis aims to propose a forecasting model to estimate the expected average profit in lending to individuals in commercial companies, thus obtaining a monetary measure to support decision making. The proposed model was developed in three major stages: 1) preprocessing, 2) classification models, and 3) model to forecast the currency risk. The first stage includes three steps: (i) delimitation of the population, (ii) sample selection, and (iii) preliminary analysis. In the second stage two more steps are necessary: (i) construction of models, and (ii) quality of the models. Finally, the last stage is regarding to the definitions for the construction of model prediction of the currency risk itself, which used the following methods: (i) ensemble, (ii) hybrid, and (iii) multiple linear regressions. The feasibility of the proposed model was tested on real data of grant credit. Results are evaluated using the prediction model in order to verify the potential increase in profits from the grant credit, comparing four scenarios: (i) without using any prevision model of credit risk, (ii) using the classification model obtained by logistic regression, (iii) using the classification model obtained with the neural network, and (iv) using the model to forecast the currency risk. The constructed model showed promising results in predicting the expected average profits, with an estimated increase of 94.97% compared to the scenario without the use of forecasting model, and an increase of 26.08% compared with the scenario of the classification model obtained by logistic regression. A sensitivity analysis of the results with variations in the profit margin per transaction was also performed, demonstrating its robustness. Accordingly, the proposed model proved effective as a support tool for management in the decision to grant credit.