António Carlos Sousa

Abstract
Resumo O abandono escolar é um problema complexo que afeta a maioria dos programas de graduação pós-secundária, em todo o mundo. O curso de engenharia industrial do Instituto ISVOUGA, localizado em Santa Maria da Feira, Portugal, não é exceção. Este estudo usou um conjunto de dados contendo informações gerais dos estudantes e suas notas para as unidades curriculares já avaliadas. A partir deste conjunto de dados, foram selecionados dezessete preditores potenciais: cinco intrínsecos (gênero, estado civil, situação profissional, idade e regime de dedicação aos estudos – integral ou parcial) e doze extrínsecos (as notas em todas as doze unidades curriculares ministradas durante os dois primeiros semestres do curso). O objetivo principal desta investigação foi prever a probabilidade de um estudante abandonar o curso com base nos referidos preditores. Foi usada uma regressão logística binária para classificar os estudantes como tendo uma probabilidade alta ou baixa de não se reinscreverem no curso. Para validar se a metodologia utilizada é apropriada para o estudo em causa, a precisão obtida com o modelo de regressão logística foi comparada, por via de uma validação cruzada com cinco partições, com a precisão obtida pela utilização de três métodos muito utilizados em data mining: One R, K Nearest Neighbors e Naive Bayes. O modelo de regressão logística identificou quatro variáveis significativas na previsão do abandono escolar (as classificações nas unidades curriculares de ciência dos materiais, eletricidade, cálculo 1 e química). Os dois preditores mais influentes do abandono dos estudantes são não conseguir aprovação nas unidades curriculares menos exigentes: ciência dos materiais e eletricidade. Ao contrário do que seria de supor antes desta investigação, descobrimos que a não aprovação em unidades curriculares mais exigentes, como física ou estatística, não tem influência significativa no abandono escolar.

Abstract
Air temperature data from many locations worldwide are only available as series of daily minima and maxima temperatures. Historically, several different approaches have been used to estimate the actual daily mean temperature, as only in the last two or three decades automatic thermometers are able to compute its actual value. The most common approach is to estimate it by averaging the daily minima and maxima. When only daily minima and maxima are available, an alternative approach, proposed by Dall'Amico and Hornsteiner in 2006, uses the two daily extremes together with next day minima temperature and a coefficient related to the local daily astronomical sunset time. Additionally, the method uses two optimizable coefficients related to the region's temperature profile. In order to use this approach it is necessary to optimize the region's unknown parameters. For this optimization, it is necessary a dataset containing the maxima, minima, and the actual daily mean temperatures for at least one year. In this research, for the period 2007-2014, we used three datasets of minima, maxima and actual mean temperatures obtained at three automatic meteorological stations located in the Douro Valley to optimize the two unknown parameters in the Dall'Amico and Hornsteiner approach. Moreover, we compared the actual mean daily temperatures available from the three datasets with the correspondent values estimated by using i) the usual approach of averaging the daily maxima and minima temperatures and ii) the Dall'Amico and Hornsteiner approach. Results show that the former approach overestimates, on average, the daily mean temperatures by 0.5 degrees C. The Dall'Amico and Hornsteiner approach showed to be a better approximation of mean temperatures for the three meteorological stations used in this research, being unbiased relative to the actual mean values of daily temperatures. In conclusion, this research confirms that the Dall'Amico and Hornsteiner is a better approach to estimate the mean daily temperatures and provides the optimized parameters for three sites located at each of the three sub-regions of the Douro Valley (Baixo Corgo, Cima Corgo and Douro Superior).

Abstract
The Douro Valley of Portugal is a well-known wine region producing Port wine since the end of the 18th century, with quality table wines becoming increasingly important over the last 20 years. Port wine production is the most important economic sector of the region and Vintage Port is the top quality Port wine type, produced only from the best vintages. The purpose of this research was to examine how the variability of annual weather influences the quality of Vintage Port. A weather and climate data set for the period 1980-2009 and a consensus ranking that combined a collection of vintage chart scores into a ranking were used to characterize both the weather and the vintage quality. In order to more precisely model the weather influences on the quality of the vintages it was necessary to partition the growing season into smaller growth intervals in which several heat and precipitation variables were evaluated. The heat-related variables were defined according to the phenology of grapevines, using a partition of the growing season based on accumulated temperature, rather than on calendar dates. Precipitation variables were calculated using broad periods corresponding to the dormant, vegetative and maturation stages of the grapevines. A logistic regression model was used as a tool to identify the weather variables that help to explain the relationships between yearly weather characteristics and vintage quality. The results show that several weather characteristics are strongly associated with better quality vintages: growing season mean temperatures above the region's average, warm winters, cool July through veraison and cool temperatures during ripening. In summary, although the weather is not solely responsible for determining a vintage quality, it plays an important role on it; therefore, its understanding can provide invaluable management insights to growers and producers.

Abstract

Abstract