Essays on Machine Learning and Hedonic Models

Thesis (Ph.D.)--University of Washington, 2016-08

Eleição nas escolas: uma análise do impacto do diretor eleito sobre o desempenho educacional no Estado da Bahia

In Brazil, the selection of school principals is set in a decentralized manner by each state and city, such that processes may vary with time for a specific locality. In the state of Bahia, school principals were appointed by a higher political hierarchy until 2008, when schools under state administration started selecting principals by elections. The main goal of this work is to evaluate whether changing this specific rule affected students proficiency levels. This is achieved by using a panel data and difference-in-differences approachs that compares state schools (treatment group) to city schools (control group) that did not face a selection rule change and thus kept having their principals politically appointed. The databases used are Prova Brasil 2007, 2009 and 2011, the first one prior and the other two former to the policy change. Our results suggest that students attending schools with principals that are selected and elected have slightly lower mean proficiency levels both in mathematics and in portuguese exams than those attending schools with appointed principals. This result, according to the literature, could be related to perverse effects of selecting school administrators by vote, such as corporatism, clientelism and politicization of the school environment

On precise three-dimensional environment modeling via UAV-based photogrammetric systems

Abstract : Images acquired from unmanned aerial vehicles (UAVs) can provide data with unprecedented spatial and temporal resolution for three-dimensional (3D) modeling. Solutions developed for this purpose are mainly operating based on photogrammetry concepts, namely UAV-Photogrammetry Systems (UAV-PS). Such systems are used in applications where both geospatial and visual information of the environment is required. These applications include, but are not limited to, natural resource management such as precision agriculture, military and police-related services such as traffic-law enforcement, precision engineering such as infrastructure inspection, and health services such as epidemic emergency management. UAV-photogrammetry systems can be differentiated based on their spatial characteristics in terms of accuracy and resolution. That is some applications, such as precision engineering, require high-resolution and high-accuracy information of the environment (e.g. 3D modeling with less than one centimeter accuracy and resolution). In other applications, lower levels of accuracy might be sufficient, (e.g. wildlife management needing few decimeters of resolution). However, even in those applications, the specific characteristics of UAV-PSs should be well considered in the steps of both system development and application in order to yield satisfying results. In this regard, this thesis presents a comprehensive review of the applications of unmanned aerial imagery, where the objective was to determine the challenges that remote-sensing applications of UAV systems currently face. This review also allowed recognizing the specific characteristics and requirements of UAV-PSs, which are mostly ignored or not thoroughly assessed in recent studies. Accordingly, the focus of the first part of this thesis is on exploring the methodological and experimental aspects of implementing a UAV-PS. The developed system was extensively evaluated for precise modeling of an open-pit gravel mine and performing volumetric-change measurements. This application was selected for two main reasons. Firstly, this case study provided a challenging environment for 3D modeling, in terms of scale changes, terrain relief variations as well as structure and texture diversities. Secondly, open-pit-mine monitoring demands high levels of accuracy, which justifies our efforts to improve the developed UAV-PS to its maximum capacities. The hardware of the system consisted of an electric-powered helicopter, a high-resolution digital camera, and an inertial navigation system. The software of the system included the in-house programs specifically designed for camera calibration, platform calibration, system integration, onboard data acquisition, flight planning and ground control point (GCP) detection. The detailed features of the system are discussed in the thesis, and solutions are proposed in order to enhance the system and its photogrammetric outputs. The accuracy of the results was evaluated under various mapping conditions, including direct georeferencing and indirect georeferencing with different numbers, distributions and types of ground control points. Additionally, the effects of imaging configuration and network stability on modeling accuracy were assessed. The second part of this thesis concentrates on improving the techniques of sparse and dense reconstruction. The proposed solutions are alternatives to traditional aerial photogrammetry techniques, properly adapted to specific characteristics of unmanned, low-altitude imagery. Firstly, a method was developed for robust sparse matching and epipolar-geometry estimation. The main achievement of this method was its capacity to handle a very high percentage of outliers (errors among corresponding points) with remarkable computational efficiency (compared to the state-of-the-art techniques). Secondly, a block bundle adjustment (BBA) strategy was proposed based on the integration of intrinsic camera calibration parameters as pseudo-observations to Gauss-Helmert model. The principal advantage of this strategy was controlling the adverse effect of unstable imaging networks and noisy image observations on the accuracy of self-calibration. The sparse implementation of this strategy was also performed, which allowed its application to data sets containing a lot of tie points. Finally, the concepts of intrinsic curves were revisited for dense stereo matching. The proposed technique could achieve a high level of accuracy and efficiency by searching only through a small fraction of the whole disparity search space as well as internally handling occlusions and matching ambiguities. These photogrammetric solutions were extensively tested using synthetic data, close-range images and the images acquired from the gravel-pit mine. Achieving absolute 3D mapping accuracy of 11±7 mm illustrated the success of this system for high-precision modeling of the environment.

A Corrupção burocrática inibe o empreendedorismo? uma análise empírica dos estados brasileiros

This work aims to investigate the relationship between the entrepreneurship and the incidence of bureaucratic corruption in the states of Brazil and Federal District. The main hypothesis of this study is that the opening of a business in Brazilian states is negatively affected by the incidence of corruption. The theoretical reference is divided into Entrepreneurship and bureaucratic corruption, with an emphasis on materialistic perspective (objectivist) of entrepreneurship and the effects of bureaucratic corruption on entrepreneurial activity. By the regression method with panel data, we estimated the models with pooled data and fixed and random effects. To measure corruption, I used the General Index of Corruption for the Brazilian states (BOLL, 2010), and to represent entrepreneurship, firm entry per capita by state. Tests (Chow, Hausman and Breusch-Pagan) indicate that the random effects model is more appropriate, and the preliminary results indicate a positive impact of bureaucratic corruption on entrepreneurial activity, contradicting the hypothesis expected and found in previous articles to Brazil, and corroborating the proposition of Dreher and Gassebner (2011) that, in countries with high regulation, bureaucratic corruption can be grease in the wheels of entrepreneurship

Financiamento do investimento em países das Américas: Uma análise de dados em painel 1972- 2008

O presente artigo tem por objetivo analisar empiricamente os determinantes do financiamento do investimento nos países das Américas no período de 1972-2008, tendo como base teórica os modelos de hiatos nas versões de Chenery e Bruno (1962); Bacha (1982, 1989 e 1990) e Taylor (1994). Com este intuito, realiza-se uma estimação com o modelo de dados em painel na versão padrão com efeitos fixos. As variáveis explicativas de cada país seguem a abordagem de Amadeo e Giambiagi (1990). São elas: os coeficientes de poupança interna e de importação e as exportações reais. Os resultados das estimações corroboram com a concepção teórica do modelo de hiatos, ou seja, que países em desenvolvimento enfrentam limitações de poupança interna e incapacidade de importar bens intensivos em capital, fatores os quais restringem o nível de investimento dessas economias.

Statistical modeling of protein lysate array data

The protein lysate array is an emerging technology for quantifying the protein concentration ratios in multiple biological samples. It is gaining popularity, and has the potential to answer questions about post-translational modifications and protein pathway relationships. Statistical inference for a parametric quantification procedure has been inadequately addressed in the literature, mainly due to two challenges: the increasing dimension of the parameter space and the need to account for dependence in the data. Each chapter of this thesis addresses one of these issues. In Chapter 1, an introduction to the protein lysate array quantification is presented, followed by the motivations and goals for this thesis work. In Chapter 2, we develop a multi-step procedure for the Sigmoidal models, ensuring consistent estimation of the concentration level with full asymptotic efficiency. The results obtained in this chapter justify inferential procedures based on large-sample approximations. Simulation studies and real data analysis are used to illustrate the performance of the proposed method in finite-samples. The multi-step procedure is simpler in both theory and computation than the single-step least squares method that has been used in current practice. In Chapter 3, we introduce a new model to account for the dependence structure of the errors by a nonlinear mixed effects model. We consider a method to approximate the maximum likelihood estimator of all the parameters. Using the simulation studies on various error structures, we show that for data with non-i.i.d. errors the proposed method leads to more accurate estimates and better confidence intervals than the existing single-step least squares method.

Accumulation, persistence and effects of indospicine residues in camels fed Indigofera plant

Indospicine (L-2-amino-6-amidinohexanoic acid) is a natural hepatotoxin found in all parts of some Indigofera plants such as I. linnaei and I. spicata. Several studies have documented a susceptibility to this hepatotoxin in different species of animals, including cattle, sheep, dogs and rats, which are associated with mild to severe liver disease after prolonged ingestion. However, there is little published data on the effects of this hepatotoxin in camels, even though Indigofera plants are known to be palatable to camels in central Australia. The secondary poisoning of dogs after prolonged dietary exposure to residual indospicine in camel muscle has raised additional food safety concerns. In this study, a feeding experiment was conducted to investigate the in vivo accumulation, excretion, distribution and histopathological effects of dietary indospicine on camels. Six young camels (2 – 4 year old), weighing 270 − 390 kg were fed daily a roughage diet consisting of Rhodes grass hay and lucerne chaff, supplemented with Indigofera and steam flaked barley. Indigofera (I. spicata) was offered at 597 mg DM/kg body weight (bw)/day designed to deliver 337 µg indospicine/kg bw/day, and fed for a period of 32 days. Blood and muscle biopsies were collected over the period of the study. Concentrations of indospicine in the plasma and muscle biopsy samples were quantitated by validated ultra-performance liquid chromatography−tandem mass spectrometry (UPLC−MS/MS). The highest concentrations in plasma (1.01 mg/L) and muscle (2.63 mg/kg fresh weight (fw)) were found at necropsy (day 33). Other tissues were also collected at necropsy and analysis showed ubiquitous distribution of indospicine, with the highest indospicine accumulation detected in the pancreas (4.86 ± 0.56 mg/kg fw) and liver (3.60 ± 1.34 mg/kg fw); followed by the muscle, heart and kidney. Histopathological examination of liver tissue showed multiple small foci of predominantly mononuclear inflammatory cells. After cessation of Indigofera intake, indospicine present in plasma in the remaining 3 camels had a longer terminal elimination half-life (18.6 days) than muscle (15.9 days), and both demonstrated mono-exponential decreases.

“Enjoy your baby” Internet-based CBT for mothers with babies: a feasibility randomised control trial

Background: Postnatal depression is a global health problem with lasting effects on the family. Government policy is focussed on early intervention and increasing access to psychological therapies. There is a growing evidence base for the use of computerised CBT packages and this study investigated the feasibility of a CBT-based self-help internet intervention for new mothers. Objective: To assess the ability to recruit mothers, deliver an internet course, obtain follow-up data and evaluate what mothers think of the course. Design: A feasibility randomised control design was used to compare a waiting list control group (delayed access= DA) to the Enjoy Your Baby course (immediate access= IA). Measures were administered at baseline and 8 week follow-up. Methods: Adverts were placed in the Metro freesheet, on charity web pages, on social media, posters were put up in the community, and leaflets were handed out at mother and baby groups. Participants had to be 18 years old or over with a child less than 18 months old. The IA arm was given access to the course straight away. After 8 weeks all participants were asked to recomplete the original measures and those in the IA arm also gave feedback on the course. Participants in the DA arm were given access after recompleting the questionnaires. Due to a lack of follow-up data a small discussion group was conducted. Intervention: The course contains 4 core modules including helping mothers understand why they feel the way they do and helping them build closeness to their babies. Additional modules, worksheets and homework tasks were available. The DA group were given a list of additional support resources and services, and encouraged to seek additional help if required. All participants received weekly automated emails for 12 weeks as they worked through the course. It was not possible to deliver individualised support. 34 Results: Despite using a number of recruitment strategies, recruitment was lower and slower than anticipated, and attrition was high. 41 women, primarily recruited via the internet, were randomised (IA n=21, DA n=20). No significant differences were observed between participants in either arm at baseline and no statistically significant differences were identified when the demographics and baseline measures of participants who logged-on to the course were compared to those who did not, or when participants who completed follow-up measures were compared to those who did not. Pre and post intervention scores on the EPDS approached statistical significance (P=.059, r=.444) favouring the intervention arm. The discussion group suggested strengths of the course and recommended areas for improvement, including making the course more mobile friendly. Conclusion: Internet interventions show promise; however it is difficult to recruit mothers, engagement is low and attrition high. A number of recommendations are made and a further pilot or an internal pilot of a larger substantive study should be conducted to confirm recruitment and retention. Trial ID: ISRCTN90927910.

Savani: A variable resolution whole-mantle model of anisotropic shear velocity variations based on multiple data sets

International audience

External school evaluation : teachers’ perspectives : two case studies from Northern Portugal

The external evaluation of non-higher education schools in Portugal has been developed by the General Inspectorate of Education since 2006. A first cycle of evaluation was completed, covering all educational units in continental Portugal up to 2011. The model of evaluation has since been subject to alterations, and a second cycle of evaluation is now coming to an end. The current model of evaluation is based on documental analysis, analysis of students’ results, and panel interviews with a variety of representatives of the school community, and addresses three domains: results, provision of educational service and management. This paper is part of an ongoing research project, developed by 6 universities and supported by the Portuguese Foundation for Science and Technology (PTDC/CPE-CED/116674/2010) which intends to analyse the impacts and effects this process of external evaluation has had on Portuguese schools. This project includes a variety of perspectives and methodologies. In particular, we will focus on two case studies undertaken in two schools from the northern region of Portugal, and more specifically on the perspectives expressed by the teachers of those schools. These particular schools were chosen because they have been evaluated twice and represent different educational levels (basic and secondary), contexts and results. These case studies included the analysis of documental data, interviews to key informants and a questionnaire directed to teachers (n = 141) – the latter will be the main focus of this paper. Teachers are essential elements of the school community when considering the impacts of external evaluation, as any changes directed at teaching practices, student evaluation, among others are only possible through their direct action and implication. Therefore, their perceptions on the process and its impacts are crucial to the understanding of what does and does not change in schools as a consequence of external evaluation. Although teachers’ opinions are not homogenous and each school reveals a number of differences when it comes to teachers’ perceptions of School Evaluation, it was possible to stress some areas as the most and as the least consensual. Teachers in both schools agree External School Evaluation (ESE) is useful for the identification of the schools’ strengths and weaknesses, values students’ external evaluation results, imposes a model for schools internal evaluation (and in fact contributes to the very existence of internal evaluation practices), and contributes to schools improvement. However teachers in both schools do not believe ESE contributes to teachers’ autonomy produces changes in how curriculum is managed, or leads to innovative teaching practices. These results point to a greater emphasis on change at the levels of school management, self-evaluation and particularly internal evaluation, but little impact on the teaching practices. We believe the classroom is at the core of school practices and teaching processes are essential to any measure of school quality and to their impacts on student learning.

Conservadorismo e herança institucional de financiamento das companhias abertas do Brasil

Tese (doutorado)—Universidade de Brasília, Universidade Federal da Paraíba, Universidade Federal do Rio Grande do Norte, Programa Multi-Institucional e Inter-Regional de Pós-Graduação em Ciências Contábeis, 2015.

A influência do conselho de administração na lucratividade das empresas listadas na BM&FBOVESPA

Dissertação (mestrado)—UnB/UFPB/UFRN, Programa MultiInstitucional e Inter-Regional de Pós-Graduação em Ciências Contábeis, 2016.

Eleição nas escolas: uma análise do impacto do diretor eleito sobre o desempenho educacional no Estado da Bahia

In Brazil, the selection of school principals is set in a decentralized manner by each state and city, such that processes may vary with time for a specific locality. In the state of Bahia, school principals were appointed by a higher political hierarchy until 2008, when schools under state administration started selecting principals by elections. The main goal of this work is to evaluate whether changing this specific rule affected students proficiency levels. This is achieved by using a panel data and difference-in-differences approachs that compares state schools (treatment group) to city schools (control group) that did not face a selection rule change and thus kept having their principals politically appointed. The databases used are Prova Brasil 2007, 2009 and 2011, the first one prior and the other two former to the policy change. Our results suggest that students attending schools with principals that are selected and elected have slightly lower mean proficiency levels both in mathematics and in portuguese exams than those attending schools with appointed principals. This result, according to the literature, could be related to perverse effects of selecting school administrators by vote, such as corporatism, clientelism and politicization of the school environment

A Corrupção burocrática inibe o empreendedorismo? uma análise empírica dos estados brasileiros

This work aims to investigate the relationship between the entrepreneurship and the incidence of bureaucratic corruption in the states of Brazil and Federal District. The main hypothesis of this study is that the opening of a business in Brazilian states is negatively affected by the incidence of corruption. The theoretical reference is divided into Entrepreneurship and bureaucratic corruption, with an emphasis on materialistic perspective (objectivist) of entrepreneurship and the effects of bureaucratic corruption on entrepreneurial activity. By the regression method with panel data, we estimated the models with pooled data and fixed and random effects. To measure corruption, I used the General Index of Corruption for the Brazilian states (BOLL, 2010), and to represent entrepreneurship, firm entry per capita by state. Tests (Chow, Hausman and Breusch-Pagan) indicate that the random effects model is more appropriate, and the preliminary results indicate a positive impact of bureaucratic corruption on entrepreneurial activity, contradicting the hypothesis expected and found in previous articles to Brazil, and corroborating the proposition of Dreher and Gassebner (2011) that, in countries with high regulation, bureaucratic corruption can be grease in the wheels of entrepreneurship

APPLICATION OF AN IMPUTATION METHOD FOR GEOSPATIAL INVENTORY OF FOREST STRUCTURAL ATTRIBUTES ACROSS MULTIPLE SPATIAL SCALES IN THE LAKE STATES, U.S.A.

Credible spatial information characterizing the structure and site quality of forests is critical to sustainable forest management and planning, especially given the increasing demands and threats to forest products and services. Forest managers and planners are required to evaluate forest conditions over a broad range of scales, contingent on operational or reporting requirements. Traditionally, forest inventory estimates are generated via a design-based approach that involves generalizing sample plot measurements to characterize an unknown population across a larger area of interest. However, field plot measurements are costly and as a consequence spatial coverage is limited. Remote sensing technologies have shown remarkable success in augmenting limited sample plot data to generate stand- and landscape-level spatial predictions of forest inventory attributes. Further enhancement of forest inventory approaches that couple field measurements with cutting edge remotely sensed and geospatial datasets are essential to sustainable forest management. We evaluated a novel Random Forest based k Nearest Neighbors (RF-kNN) imputation approach to couple remote sensing and geospatial data with field inventory collected by different sampling methods to generate forest inventory information across large spatial extents. The forest inventory data collected by the FIA program of US Forest Service was integrated with optical remote sensing and other geospatial datasets to produce biomass distribution maps for a part of the Lake States and species-specific site index maps for the entire Lake State. Targeting small-area application of the state-of-art remote sensing, LiDAR (light detection and ranging) data was integrated with the field data collected by an inexpensive method, called variable plot sampling, in the Ford Forest of Michigan Tech to derive standing volume map in a cost-effective way. The outputs of the RF-kNN imputation were compared with independent validation datasets and extant map products based on different sampling and modeling strategies. The RF-kNN modeling approach was found to be very effective, especially for large-area estimation, and produced results statistically equivalent to the field observations or the estimates derived from secondary data sources. The models are useful to resource managers for operational and strategic purposes.