Doctor of Philosophy / Department of Statistics / Weixing Song / Kernel based non-parametric regression is a popular statistical tool to identify the relationship between response and predictor variables when standard parametric regression models are not appropriate. The efficacy of kernel based methods depend both on the kernel choice and the smoothing parameter. With insufficient smoothing, the resulting regression estimate is too rough and with excessive smoothing, important features of the underlying relationship is lost. While the choice of the kernel has been shown to have less of an effect on the quality of regression estimate, it is important to choose kernels to best match the support set of the underlying predictor variables. In the past few decades, there have been multiple efforts to quantify the properties of asymmetric kernel density and regression estimators. Unlike classic symmetric kernel based estimators, asymmetric kernels do not suffer from boundary problems. For example, Beta kernel estimates are especially suitable for investigating the distribution structure of predictor variables with compact support. In this dissertation, two types of Beta kernel based non parametric regression estimators are proposed and analyzed. First, a Nadaraya-Watson type Beta kernel estimator is introduced within the regression setup followed by a local linear regression estimator based on Beta kernels. For both these regression estimators, a comprehensive analysis of its large sample properties is presented. Specifically, for the first time, the asymptotic normality and the uniform almost sure convergence results for the new estimators are established. Additionally, general guidelines for bandwidth selection is provided. The finite sample performance of the proposed estimator is evaluated via both a simulation study and a real data application. The results presented and validated in this dissertation help advance the understanding and use of Beta kernel based methods in other non-parametric regression applications.

The characteristics of vehicle steering perception are decisive factors concerning vehicle safety and overall pleasure behind the wheel. It is a challenge for vehicle manufacturers to achieve these features and qualities, because usually vehicle tuning almost only relies on subjective evaluation of test drivers, which is costly and time consuming. In order to optimize suspension design and develop a tool that can be used to evaluate steering with objective metrics instead of subjective assessment, links between them must be confirmed. In this master thesis, both objective and subjective testing data of over 20 vehicles across four different segments are introduced in linear and nonlinear analysis. Linear regression analysis is applied to investigate simply positive or negative correlation between a pair of subjective-objective parameters. However, even if certain linear correlations are obtained, it is still hard to define the optimal value for objective metrics. Considering that the general shape of a correlation function can reveal which objective range give higher subjective rating, it is possible to define these preferred ranges with Neural Network (NN). The best data available is adopted from three drivers who tested 15 sedans, and some interesting results are found. The initial results demonstrate that NN is a powerful tool to uncover and graphically illustrate the links between objective metrics and subjective assessments, i.e., the specific range leading to better steering feel. Given a larger sample size, more reliable and optimal links can be defined by following the same method. These confirmed links would enable vehicle dynamics engineers to more effectively develop new vehicles with nearly perfect steering feel.

The load forecasting has become an important role in the operation of power system, and several models by using different techniques have been applied to solve these problems. In the literature, the linear regression models are considered as a traditional approach to predict power consumption, and more recently, the artificial neural network (ANN) models have received more attention for a great number of successful and practical applications. This report introduces both linear regression and ANN models to predict the power consumption for Fortum in Ekerö. The characteristics of power consumption of different kinds of consumers are analyzed, together with the effects of weather parameters to power consumption. Further, based on the gained information, the numerical models of load forecasting are built and tested by the historical data. The predictions of power consumption are focus on three cases separately: total power consumption in one year, daily peak power consumption during winter and hourly power consumption. The processes of development of the models will be described, such as the choice of the variables, the transformations of the variables, the structure of the models and the training cases of ANN model. In addition, two linear regression models will be built according to the number of input variables. They are simple linear regression with one input variable and multiple linear regression with several input variables. Comparison between the linear regression and ANN models will be carried out. In the end, it finds out that the linear regression obtains better results for all the cases in Ekerö. Especially, the simple linear regression outperforms in prediction of total power consumption in one year, and the multiple linear regression is better in prediction of daily peak load during the winter.

Master of Science / Department of Statistics / Weixin Yao / In this report, we investigate a robust estimation of the number of components in the mixture of regression models using trimmed information criterion. Compared to the traditional information criterion, the trimmed criterion is robust and not sensitive to outliers. The superiority of the trimmed methods in comparison with the traditional information criterion methods is illustrated through a simulation study. A real data application is also used to illustrate the effectiveness of the trimmed model selection methods.

Résumé: La mesure de l’équivalent en eau de la neige (EEN) sur le terrain permet de prédire la quantité d’eau libérée par la fonte de la neige. La télédétection dans les micro-ondes passives offre le potentiel d’estimer I’EEN et peut complémenter ces observations de façon synoptique pour l’ensemble du territoire. Un produit de cartographie de I’EEN couvrant l’ensemble du globe a été élaboré par le NSIDC basé sur le capteur AMSR-E. Cet instrument, lancé en 2002, a une résolution améliorée par rapport aux capteurs antérieurs. L’estimation de I’EEN se base sur la différence entre un canal peu affecté (19 GHz) et un canal affecté (37 GHz) par la diffusion de volume de la neige. La précision de ce produit a été évaluée pour la province de Québec à l’hiver 2003 et à l’hiver 2004 qui ont un EEN moyen de 170 mm. Des sous-estimations importantes ont été révélées et une certaine difficulté à détecter la présence de neige. Des modèles régionaux de régressions linéaires ont été développés pour le Québec. Des corrections pour la fraction d’eau et de forêt ont été appliquées à la combinaison T19v.37v et ont permis d’améliorer les résultats. Ces corrections sont basées sur la température de l’air du modèle GEM. Les meilleurs résultats sont pour la classe de neige taïga à l’hiver 2003 avec une erreur relative de 24 % tandis que l’erreur relative est d’environ 40 % pour la région maritime. Les erreurs élevées dans la classe taïga ont été attribuées à des couverts de neige plus épais que la capacité de pénétration des micro-ondes tandis que les erreurs de la classe maritime a des fractions forêt élevées et à la neige mouillée. La présence d’importante quantité de neige et la forêt dense de la province de Québec compliquent l’estimation de I’EEN au Québec avec un modèle de régression. || Abstract: Snow water equivalent (SWE) measurements in the field allow estimation of the quantity of released water from the melting of snow. This is useful to predict the water reserve available for production of hydro-electricity. Remote sensing with microwave can estimate SWE and complement those observations synoptically for whole territories. A SWE mapping products was developed by NSIDC based on the AMSR-E sensor launched in 2002 with an improved resolution compared to previous sensors. SWE estimation is based on difference between a channel weakly affected (19 GHz) and a channel strongly affected by volume scattering. The precision of this product was evaluated for the province of Quebec in winter 2003 and winter 2004 with a mean SWE of 170 mm. Important underestimation and some difficulty of detecting the snow was revealed. Regional linear regression models were developed for the province of Quebec. Corrections for forest and water fraction were applied on T19V-37V combination and permit to improve the results. Those corrections were based on air temperature from the GEM model. Best results were found for taiga snow class in winter 2003 with a relative error of 28% and approximately 40% for maritime snow class. High errors in the taiga region were attributed to snow depth higher than the penetration depth of the microwave and errors in the maritime region to high forest density and wet snow. The important snow amount and high density forest of the province of Quebec hampers the estimation of SWE with a regression model.

Background: The Active Commuting Route Environment Scale (ACRES) was developed to study active commuters’ perceptions of their route environments. The overall aims were to assess the measuring properties of the ACRES and study active bicycle commuters’ perceptions of their commuting route environments. Methods: Advertisement- and street-recruited bicycle commuters from Greater Stockholm, Sweden, responded to the ACRES. Expected differences between inner urban and suburban route environments were used to assess criterion-related validity, together with ratings from an assembled expert panel as well as existing objective measures. Reliability was assessed as test-retest reproducibility. Comparisons of ratings between advertisement- and street-recruited participants were used for assessments of representativity. Ratings of inner urban and suburban route environments were used to evaluate commuting route environment profiles. Simultaneous multiple linear regression analyses were used to assess the relation between the outcome variable: whether the route environment hinders or stimulates bicycle-commuting and environmental predictors, such as levels of exhaust fumes, speeds of traffic and greenery, in inner urban areas. Results: The ACRES was characterized by considerable criterion-related validity and reasonable test-retest reproducibility. There was a good correspondence between the advertisement- and street-recruited participants’ ratings. Distinct differences in commuting route environment profiles between the inner urban and suburban areas were noted. Suburban route environments were rated as safer and more stimulating for bicycle-commuting. Beautiful, green and safe route environments seem to be, independently of each other, stimulating factors for bicycle-commuting in inner urban areas. On the other hand, high levels of exhaust fumes and traffic congestion, as well as low ‘directness’ of the route, seem to be hindering factors. Conclusions: The ACRES is useful for assessing bicyclists’ perceptions of their route environments. A number of environmental factors related to the route appear to be stimulating or hindering for bicycle commuting. The overall results demonstrate a complex research area at the beginning of exploration. / BAKGRUND: Färdvägsmiljöer kan tänkas påverka människors fysiskt aktiva arbetspendling och därmed bidra till bättre folkhälsa. Studier av färdvägsmiljöer är därför önskvärda för att öka förståelsen kring möjliga samband mellan fysiskt aktiv arbetspendling och färdvägsmiljöer. En enkät, ”The Active Commuting Route Environment Scale” (ACRES), har därför skapats i syfte att studera fysiskt aktiva arbetspendlares upplevelser av sina färdvägsmiljöer. Huvudsyftet med denna avhandling var dels att studera enkätens psykometriska egenskaper i form av validitet och reliabilitet, dels att studera arbetspendlande cyklisters upplevelser av sina färdvägsmiljöer. METODER: Arbetspendlande cyklister från Stor-Stockholm rekryterades via tidningsannonsering och via direkt kontakt i anslutning till färdvägen. Deltagarna besvarade enkäten ACRES. Tillsammans med skattningar från en grupp av experter och redan existerande objektiva mått användes förväntade skillnader mellan färdvägsmiljöer i inner- och ytterstaden för att studera kriterierelaterad validitet. Reliabiliteten studerades som reproducerbarhet via upprepade mätningar (test-retest). Jämförelser mellan skattningar av deltagare rekryterade via annonsering och via direkt kontakt i färdvägsmiljöer användes för att studera representativitet. Skattningar av färdvägsmiljöer i inner- och ytterstaden användes vidare för att studera färdvägsmiljöprofiler. Multipel linjär regressionsanalys användes även för att studera sambandet mellan utfallsvariabeln huruvida färdvägsmiljön motverkar eller stimulerar arbetspendling med cykel och miljöprediktorer, såsom avgasnivåer, trafikens hastighet och grönska, i innerstadsmiljöer. RESULTAT: Enkäten ACRES visade god kriterierelaterad validitet och rimlig reproducerbarhet. Det var en god överrensstämmelse mellan skattningar av deltagare rekryterade via annonsering och via direkt kontakt. Färdvägsmiljöprofilerna visade tydliga skillnader mellan inner- och ytterstadsmiljöer. Ytterstadens färdvägsmiljöer skattades som tryggare och mer stimulerande för arbetspendling med cykel än innerstadens färdvägsmiljöer. Vidare verkar vackra, gröna och trygga färdvägsmiljöer, oberoende av varandra, vara stimulerade faktorer för arbetspendling med cykel i innerstadsmiljöer. Däremot verkar höga avgasnivåer, höga trängselnivåer och färdvägar som kräver många riktningsändringar vara motverkande faktorer. SLUTSATSER: Enkäten ACRES är ett användbart instrument vid mätningar av cyklisters upplevelser av sina färdvägsmiljöer. Ett antal faktorer relaterade till färdvägsmiljön verkar vara stimulerande respektive motverkande för arbetspendling med cykel. Generellt sett på visar resultaten ett relativt outforskat och komplext forskningsområde. / <p>Örebro universitet, Hälsoakademin</p> / FAAP

Coconuts are one of the most economically important plants in Mozambique, where millions of people depend on income from coconuts. The coconut lethal yellowing disease (CLYD) is a highly destructive disease that ever since the early 90’s causes coconut palms in Mozambique to stop producing fruit and leave the coconut farmers with only empty stems. This thesis examines the disease's effect on the farmers’ income, both from coconuts and other complementary sources, since the vendible harvest should decrease with the incidence of the disease. The method used is multivariate linear regression, where several income variables are used as dependent variables. Two models are created, one only interpreted for the sample of 488 observations and one aiming at generalizing the results. By this study, it cannot be confirmed that the incidence of CLYD has a significant effect on coconut farmers’ income. The results from the sample analysis do however show that the income is affected by the degree of the disease, which is an incentive for continued research in the field.

A fundamental question on visual system in neuroscience is how the visual stimuli are functionally related to neural responses. This relationship is often explained by the notion of receptive fields, an approximated linear or quasi-linear filter that encodes the high dimensional visual stimuli into neural spikes. Traditional methods for estimating the filter do not efficiently exploit prior information about the structure of neural receptive fields. Here, we propose several approaches to design the prior distribution over the filter, considering the neurophysiological fact that receptive fields tend to be localized both in space-time and spatio-temporal frequency domain. To automatically regularize the estimation of neural receptive fields, we use the evidence optimization technique, a MAP (maximum a posteriori) estimation under a prior distribution whose parameters are set by maximizing the marginal likelihood. Simulation results show that the proposed methods can estimate the receptive field using datasets that are tens to hundreds of times smaller than those required by traditional methods. / text

Privacy preserving distributed data mining aims to design secure protocols which allow multiple parties to conduct collaborative data mining while protecting the data privacy. My research focuses on the design and implementation of privacy preserving two-party protocols based on homomorphic encryption. I present new results in this area, including new secure protocols for basic operations and two fundamental privacy preserving data mining protocols. I propose a number of secure protocols for basic operations in the additive secret-sharing scheme based on homomorphic encryption. I derive a basic relationship between a secret number and its shares, with which we develop efficient secure comparison and secure division with public divisor protocols. I also design a secure inverse square root protocol based on Newton's iterative method and hence propose a solution for the secure square root problem. In addition, we propose a secure exponential protocol based on Taylor series expansions. All these protocols are implemented using secure multiplication and can be used to develop privacy preserving distributed data mining protocols. In particular, I develop efficient privacy preserving protocols for two fundamental data mining tasks: multiple linear regression and EM clustering. Both protocols work for arbitrarily partitioned datasets. The two-party privacy preserving linear regression protocol is provably secure in the semi-honest model, and the EM clustering protocol discloses only the number of iterations. I provide a proof-of-concept implementation of these protocols in C++, based on the Paillier cryptosystem.

Community and Regional Planning / In the past two decades, Austin has tremendous population growth, job opportunity in the downtown core and transportation challenges associated with that. Public transit, and particularly rail, often is regarded as a strategy to help reduce urban traffic congestion. The Urban Rail, which combines features of streetcars and light rail, is introduced into Austin as a new transit rail. The City of Austin, Capital Metro and Lone Star Rail are actively studying routing, financial, environmental and community elements associated with a first phase of Urban Rail. This thesis collected 2010 Origin and Destination Rail Transit Survey data from Capital Metropolitan Transportation Authority. The research focuses on the rail transit ridership. Two regression models are applied to analyze the factors influencing Austin rail transit ridership. One model is focusing on the socioeconomic characteristics. One model is focusing on the spatial factors. Our model shows that demographic factors have more significant effect than spatial factors. In addition, this work also tries to analyze the correlations between those factors and make recommendations based on the analysis result. / text