Dissertação (mestrado)—Universidade de Brasília, Faculdade de Economia, Administração, Contabilidade e Gestão Pública, Programa de Pós-Graduação em Administração, 2016. / Submitted by Fernanda Percia França (fernandafranca@bce.unb.br) on 2017-03-29T16:54:14Z No. of bitstreams: 1 2016_PengYaohao.pdf: 1180450 bytes, checksum: ef8f0884103e32b6a4dac98a1c9dd880 (MD5) / Approved for entry into archive by Raquel Viana(raquelviana@bce.unb.br) on 2017-04-13T20:57:38Z (GMT) No. of bitstreams: 1 2016_PengYaohao.pdf: 1180450 bytes, checksum: ef8f0884103e32b6a4dac98a1c9dd880 (MD5) / Made available in DSpace on 2017-04-13T20:57:38Z (GMT). No. of bitstreams: 1 2016_PengYaohao.pdf: 1180450 bytes, checksum: ef8f0884103e32b6a4dac98a1c9dd880 (MD5) / O presente estudo realizou a previsão da taxa spot de 15 pares de câmbio mediante a aplicação de um algoritmo de aprendizado de máquinas – Support Vector Regression – com base em um modelo fundamentalista composto por 13 variáveis explicativas. Para a estimação das previsões, foram consideradas 9 funções Kernel extraídas da literatura científica, totalizando assim 135 modelos verificados. As previsões foram comparadas com o benchmark Random Walke avaliadas em relação à taxa de acerto direcional do câmbio e às métricas de erro RMSE (raiz quadrada do erro quadrático médio) e MAE (erro absoluto médio). A significância estatística do incremento de poder explicativo dos modelos SVR em relação ao Random Walk foi verificada mediante a aplicação do Reality Check Test de White (2000). Os resultados mostram que os modelos SVR obtiveram desempenho preditivo satisfatório em relação ao benchmark, com vários dos modelos propostos apresentando forte significância estatística de superioridade preditiva.Por outro lado, observou-se que várias funções Kernel comumente utilizadas na literatura científica não lograram êxito em superar o Random Walk, apontando para uma possível lacuna no estado da arte de aprendizado de máquinas aplicada à previsão de taxas de câmbio. Por fim, discutiu-se acerca das implicações dos resultados obtidos para o desenvolvimento futuro da agenda de pesquisa correlata. / This paper aims to forecast the spot exchange rate of 15 currency pairs by applying a machinelearning algorithm – Support Vector Regression – based on a fundamentalist model composedof 13 explanatory variables. The predictions’ estimation were obtained by applying 9different Kernel functions extracted from the scientific literature, resulting in a total of 135 modelsverified. The predictions were compared to the Random Walk benchmark and evaluated for directionalaccuracy rate of exchange pradictions and error performance indices RMSE (root meansquare error) and MAE (mean absolute error). The statistical significance of the explanatorypower gain via SVR models with respect to the Random Walk was checked by applying White(2000)’s Reality Check Test. The results show that SVR models achieved satisfactory predictiveperformance relative to the benchmark, with several of the proposed models showing strong statisticalsignificance of predictive superiority. Furthermore, the results showed that mainstreamKernel functions commonly used in the scientific literature failed to outperform the RandomWalk,indicating a possible gap in the state of art of machine learning methods applications to exchangerates forecasting. Finally, the paper presents a discussion about the implications of the obtainedresults for the future development of related research agendas.

abstract: Background: Twenty-four hour urinary sucrose and fructose (24uSF) has been developed as a dietary biomarker for total sugars intake. Collection of 24-h urine is associated with high costs and heavy participant burden, while collection of spot urine samples can be easily implemented in research protocols. The aim of this thesis is to investigate the utility of uSF biomarker measured in spot urine. Methods: 15 participants age 22 to 49 years completed a 15-day feeding study in which they consumed their usual diet under controlled conditions, and recorded the time each meal was consumed. Two nonconsecutive 24-hour urines, where each urine void was collected in a separate container, were collected. Four timed voids (morning, afternoon, evening, and next day) were identified based on time of void and meal time. Urine samples were measured for sucrose, fructose and creatinine. Variability of uSF excretion was assessed by coefficient of variation (%CV) and variance ratios. Pearson correlation coefficient and multiple linear regression were used to investigate the association between uSF in each timed void and corresponding 24uSF excretion. Results: The two-day mean uSF was 50.6 mg (SD=29.5) for the 24-h urine, and ranged from 4.5 to 7.5 mg/void for the timed voids. The afternoon void uSF had the lowest within-subject variability (49.1%), and lowest within- to between-subject variance ratio (0.2). The morning and afternoon void uSF had the strongest correlation with 24-h uSF for both mg/void (r=0.80 and r=0.72) and mg/creatinine (r=0.72 and r=0.67), respectively. Finally, the afternoon void uSF along with other covariates had the strongest predictive ability of 24-h uSF excretion (mg/void) (Adjusted R2= 0.69; p=0.002), whereas the morning void had the strongest predictive ability of 24-h uSF excretion (mg/g creatinine) (adjusted R2= 0.58; p=0.008). Conclusions: The afternoon void uSF had the most favorable reproducibility estimates, strong correlation with 24uSF excretion, and explained greatest proportion of the variability in 24uSF. USF in mg/void may be better to use than uSF in mg/g creatinine as a biomarker in spot urine. These findings need to be confirmed in a larger study, and in a study population with a wide range of sugars intake. / Dissertation/Thesis / Masters Thesis Nutrition 2018

Dollar spot, caused by the pathogen S. homoeocarpa (F.T. Bennett), is a common disease that infects a wide variety of turfgrasses all over the world. Yet it is significant problem on golf course putting greens and fairways consisting of creeping bentgrass (Agrostis stolonifera L.) and annual bluegrass (Poa annua L.). It is active in a wide variety of environmental conditions ranging from 16-30˚C but favors warm, humid days, followed by cool nights. Sclerotinia homoeocarpa overwinters as dormant mycelium in dead plant tissue. In the spring, germinating mycelia begin to infect leaf blades causing foliar lesions, which then spread via mycelium by means of wind, rain, animals and equipment. While there are a number of cultural practices that can reduce disease severity, frequent fungicide applications are required to maintain acceptable playing conditions on a golf course. The repeated use of fungicides with the same mode of action has led to the development of fungicide resistance of S. homoeocarpa to certain fungicide classes. Most notably, demethylase inhibitor (DMI) fungicides have been found to have varying levels of inefficacy against S. homoeocarpa across North America. The cause for reduced efficacy is suspected to the shifted sensitivity levels of many S. homoeocarpa populations, which are resulted from repeated use of the DMI fungicide. Recently, “early-spring fungicide applications” targeting to reduce initial inoculum density of dollar spot have gained popularity in an attempt to reduce dollar spot severity. In addition, preventative fungicide applications (from late October through mid-November) containing DMI fungicides have been traditionally practiced to target snow molds (caused by Microdochium nivale, Typhula spp.) in the northeastern United States. To date, there is not a clear understanding as to what effect, if any, these applications have on S. homoeocarpa DMI sensitivity or residual dollar spot control the following year. Traditional preventative snow mold applications were also investigated on the effect of S. homoeocarpa DMI sensitivity and early-season dollar spot control. The objective of this study was to investigate the effect of early-spring dollar spot application and late-fall snow mold application on S. homoeocarpa population with a bimodal distribution of DMI sensitive and insensitive isolates.

Frogeye leaf spot, caused by Cercospora sojina Hara, is a foliar disease affecting soybean (Glycine max (L.) Merr.), often managed by applications of quinone outside inhibitor (QoI) fungicides. In 2013 and 2014, symptomatic leaf samples were collected from Mississippi soybean fields leading to the collection of 634 mono-conidial C. sojina isolates. In vitro bioassays were performed to evaluate the sensitivity of 14 isolates plus a baseline. Resistant and sensitive isolates were characterized by determining the effective fungicide concentrations at which 50% of conidial germination was inhibited (EC50). Additionally, the molecular mechanism of resistance was determined for all 634 isolates. Greater than 93% of C. sojina isolates collected in Mississippi carried the G143A amino acid substitution indicating a shift to a QoI-resistant population throughout Mississippi soybean fields. Greenhouse studies confirmed that due to this amino acid substitution, symptoms caused by QoI-resistant isolates developed in spite of a QoI fungicide application.

Soybean is an economically important crop. It is a selfertilized species grown on vast contiguous acres. These facts predispose soybean to disease epidemics. Cercospora sojina, causal agent of frogeye leaf spot, has reduced United States soybean productivity 0.3 percent on average per year between 2008 and 2010. Several states have reported the pathogen developing resistance to the strobilurin class of fungicides. To date genetic host resistance has been identified as single dominant genes (Rcs1, Rcs2, and Rcs3). However, the lifespans of Rcs1 and Rcs2 were 10 and 16 years respectively. Currently, the Rcs3 locus has been utilized in all major soybean breeding programs of the US and has been for over 20 years. Seventyive accessions of soybean were found to exhibit resistance to multiple races of C. sojina while not exhibiting the Rcs3 haplotype. Twenty of these plant introductions (PIs) were screened by six races within the new race classification system of C. sojina representing all domestic variability of the pathogen. Two agronomically favorable PIs, PI398993 and PI399068, were found in this research to exhibit broad resistance to sources documented to contain most domestic variability of the pathogen. Two segregating populations were developed by crossing PI398993 x ‘Blackhawk’ and PI399068 x Blackhawk. Segregation ratios of F2 as well as F2:3 family seedling screens of both populations indicating single dominant gene action in both resistance sources. Single marker analysis indicated markers associated with the phenotype were indeed on chromosome 16 (MLG J), but possibly beyond Rcs3 in both sources. Interval mapping placed the highest probability of the resistance loci near SNP_171 and SNP_368, 72.86 and 72.48 cM respectively, but distal to the Rcs3 locus. Analysis of reaction ratings also indicated significant influence on phenotype was also associated with markers located at or beyond the published Rcs3 locus. The evidence in this research supports the hypothesis that both PIs may contain a resistance loci, potentially different than Davis, but within the same gene cluster. Equally as likely, the resistance could prove allelic to Davis.

Spot friction welding (SFW) was performed on ultra high strength steel (UHSS) steel sheet commonly used in automobile manufacturing. Alloys studied included DP780, DP780EG, DP980, and DF140T sheet steel of varying thickness from 1.2 mm to 1.4 mm. Welding was accomplished using a PCBN standard tool. Weld strengths were then compared to a proposed AWS standard. Initial hardness readings were taken in cross sectioned samples. Grain structure in a SFW is presented. Resistance spot welds were created in three steels. This study focuses on the strength of SFW joints as compared to traditional resistance spot welding (RSW) in welding like materials to one another. Cycle times of SFW were also evaluated and compared to production rate cycle times of RSW.

Refill friction stir spot welding (RFSSW) is an emerging technology, capable of joining thin sheets of aluminum alloys. The present thesis comprises two studies which were conducted to address two challenges faced by RFSSW: the long cycle time traditionally associated with welding and the poor life of existing RFFSW tools. In the first study, welds were made in AA5052-H36, at various cycle times and with various process parameters. It was shown that RPM, cycle time, and material thickness, all have an effect on the machine response. Decreasing RPM or weld duration leads to increased force and torque response during welding. Welds with cycle times below one second were successfully made without severely impacting joint quality, suggesting that prior work may have been limited by machine capabilities rather than by phenomena inherent to the process. On average, the sub-one second welds caused a peak probe force of 9.81 kN, a plunge torque of 26.3 N*m, and showed average lap-shear strengths of 7.0kN; compared to a peak probe force of 5.14 kN, a plunge torque of 17.3 N*m, and lap-shear strength of 6.89kN for a more traditional four-second welding condition. In the second study, the life of a steel toolset was quantified as consecutive welds were made in AA5052-H36 until the toolset seized from material accumulation/growth. At a one-second welding condition, the toolset was only capable of producing 53 welds before seizure. At a two-second welding condition, the toolset was only capable of producing 48 welds. In subsequent temperature experiments, thermocouples were embedded into welding coupons at various locations near weld center, allowing novel temperature data to be collected for welds with varying cycle times and parameters. The collected temperature data shows that as cycle time increases, so does weld temperature. At weld center, temperatures in excess of 500°C were observed in welds with 4 second durations. At these temperatures, Fe-Al intermetallic growth is anticipated as a mechanism limiting the tool life observed. The results suggest that steel is not an appropriate choice for RFSSW tools, and future evaluation of other materials is merited.

An acousto-ultrasonic approach is used to explore the damage development in tensile shear spot welds during fatigue loading. There is reasonable data to support the hypothesis that a decrease in an AU signal is indicative of the presence of an internal crack and could be used for monitoring and evaluation purposes. / Master of Science

The market for natural gas in the United States follows a yearly price pattern of high prices during the winter heating season and lows during the summer months. During the winter heating season the daily and weekly price fluctuations for natural gas are normally related to ambient air temperature and other weather related phenomenon. This paper examines a natural gas price forecasting model developed by the U.S. Department of Energy, Energy Information Agency (EIA). This paper proposes that a more accurate forecasting model can be created from the EIA model by focusing on forecasting price during only the winter heating season and by adding other variables to the EIA model. The forecasting results of the core EIA model are compared to the results of other linear regression models. / Master of Arts

This study computed the Normalized Difference Vegetation Index (NDVI) products derived from NOAA AVHRR, MODIS, and SPOT VGT sensors. NDVI products from different instruments vary in spatial resolution, temporal coverage and spectral range. As a result, multi-sensor NDVI products are rarely used in a single phenological study. In order to evaluate the difference and similarity of NDVI records from the three sensors, I used EPA Eco-region frameworks to determine the average annual Start of Season (SOS) and End of Season (EOS) of Contiguous United States, and analyzed dates among datasets. In addition, I created 1127 sample points within the study area, and compared relationship between SOS/EOS based on land cover. The objectives of this thesis are to: 1) compare multi-sensor NDVI data using phenological models, 2) define a strategy to merge multi-sensor NDVI products to a single phenological product without direct NDVI conversion. The spatial and statistical analysis revealed that the Land Surface Phenology (LSP) measurements retrieved from NDVI time series from different sensors follow linear and positive relationships where compared by either eco-region or sample point. The historical record of AVHRR combined with the modern MODIS and SPOT data provides a critical and reliable perspective on phenological patterns in Contiguous United States area. The success of this study will help LSP by providing understanding of how different instruments can be combined to generate multi-sensor NDVI data for phenology. / Master of Science