The classical Poisson, geometric and negative binomial regression models for count data belong to the family of generalized linear models and are available at the core of the statistics toolbox in the R system for statistical computing. After reviewing the conceptual and computational features of these methods, a new implementation of zero-inflated and hurdle regression models in the functions zeroinfl() and hurdle() from the package pscl is introduced. It re-uses design and functionality of the basic R functions just as the underlying conceptual tools extend the classical models. Both model classes are able to incorporate over-dispersion and excess zeros - two problems that typically occur in count data sets in economics and the social and political sciences - better than their classical counterparts. Using cross-section data on the demand for medical care, it is illustrated how the classical as well as the zero-augmented models can be fitted, inspected and tested in practice. (author's abstract) / Series: Research Report Series / Department of Statistics and Mathematics

The classical Poisson, geometric and negative binomial regression models for count data belong to the family of generalized linear models and are available at the core of the statistics toolbox in the R system for statistical computing. After reviewing the conceptual and computational features of these methods, a new implementation of hurdle and zero-inflated regression models in the functions hurdle() and zeroinfl() from the package pscl is introduced. It re-uses design and functionality of the basic R functions just as the underlying conceptual tools extend the classical models. Both hurdle and zero-inflated model, are able to incorporate over-dispersion and excess zeros-two problems that typically occur in count data sets in economics and the social sciences-better than their classical counterparts. Using cross-section data on the demand for medical care, it is illustrated how the classical as well as the zero-augmented models can be fitted, inspected and tested in practice. (authors' abstract)

This dissertation identifies key determinants of outbound cross-border mergers and acquisitions (M&As) by emerging Asian acquirers during 2001-2012. Using a zero-inflated model that takes into account different mechanisms governing country pairs that never engage in cross-border M&As and country pairs that actively participate in cross-border M&As, I uncover unique patterns for emerging Asian acquirers. Emerging Asian acquirers originate from countries with lower corporate tax rates than those countries where their targets are located. Furthermore, the negative impact of an international double tax burden is significantly larger than that found in previous studies. While country governance differences and geographical and cultural differences are important determinants of international M&As, relative valuation effects are muted. Coefficients of these determinants vary substantially, depending on whether targets are located in developing or advanced nations. Also, determinants differ considerably between active and non-active players in cross-border M&As. Moreover, comparisons of empirical models illustrate that estimating a non-linear model and taking into account both the bounded nature and non-normal distributions of fractional response variables lead to different inferences from those drawn from a linear model estimated by the ordinary least squares method. Overall, emerging Asian acquirers approach the deals differently from patterns documented in developed markets. So, when evaluating foreign business combinations or devising policies, managers or policymakers should consider these differences.

The use of Light Detection and Ranging (LiDAR) information is gaining popularity, however its use has been limited in deciduous forests. This thesis describes two studies using LiDAR data in an Eastern Kentucky deciduous forest. The first study quantifies vertical error of LiDAR derived digital elevation models (DEMs) which describe the forests terrain. The study uses a new method which eliminates Global Positioning System (GPS) error. The study found that slope and slope variability both significantly affect DEM error and should be taken in to account when using LiDAR derived DEMs. The second study uses LiDAR derived forest vegetation and terrain metrics to predict terrestrial Plethodontid salamander abundance across the forest. This study used night time visual encounter surveys coupled with zero-inflation modeling to predict salamander abundance based on environmental covariates. We focused on two salamander species, Plethodon glutinosus and Plethodon kentucki. Our methods produced two different best fit models for the two species. Plethodon glutinosus included vegetation height standard deviation and water flow accumulation covariates, while Plethodon kentucki included only canopy cover as a covariate. These methods are applicable to many different species and can be very useful for focusing management efforts and understanding species distributions across the landscape.

In this study, a panel data set of flights made by employees at the Royal Institute of Technology (KTH) in Sweden is analyzed using generalized linear modeling approaches, with the aim to create a model with high predictive capability of the quarterly CO2 emission and the number of flights, for a year not included in the model estimation. A Zero-inflated Gamma regression model is fitted to the CO2 emission variable and a Zero-inflated Negative Binomial regression model is used for the number of flights. To build the models, cross-validation is performed with the observations from 2018 as the training set and the observations from the next year, 2019, as the test set. One at a time, the variable that best improves the prediction of the test set data (either as included in the count model or the zero-inflation model) is selected until an additional variable turns out insignificant on a 5% significance level in the estimated model. In addition to the variables in the data, three lags of the dependent variables (CO2 emission and flights) were included, as well as transformed versions of the continuous variables, and a random intercept each for the categorical variables indicating quarter and department at KTH, respectively. Neither model selected through the cross-validation process turned out to be particularly good at predicting the values for the upcoming year, but a number of variables were proven to have a statistically significant association with the respective dependent variable.

Dans la première partie de cette thèse, nous avons mis en place un métamodèle de transmission du paludisme basé sur la modélisation compartimentale susceptible-infecté-résistant (SIR) et prenant en compte les flux de mobilité humaine entre différents villages du Centre Sénégal. Les stratégies d’intervention géographiquement ciblées, s’étaient avérées efficaces pour réduire l’incidence du paludisme aussi bien dans les zones d’intervention qu’à l’extérieur de ces zones. Cependant, des actions combinées ciblant à la fois le vecteur et l’hôte, coordonnées à large échelle sont nécessaires dans les régions et pays visant l’élimination du paludisme à court/moyen terme.Dans la deuxième partie nous avons évalué différentes méthodes d’estimation de la mobilité humaine en l’absence de données individuelles. Ces méthodes incluaient la traçabilité spatio-temporelle des téléphones mobiles ainsi que les modèles mathématiques de gravité et de radiation. Le transport de l’agent pathogène dans l’espace géographique, par la mobilité d’un sujet infecté est un déterminant majeur de la vitesse de propagation d’une épidémie. Nous avons introduit le modèle d’impédance qui minimise l’erreur quadratique moyen sur les estimations de mobilité, en particulier dans les contextes où les ensembles de population sont caractérisés par leurs tailles hétérogènes.Nous avons enfin élargi le cadre des hypothèses sous-jacentes à la calibration des modèles de gravité de la mobilité humaine. L’hypothèse d’une distribution avec excès de zéros a fourni un meilleur ajustement et une meilleure prédictibilité, comparée aux hypothèses classiques n’assumant pas un excès de zéros : Poisson, Quasipoisson. / In the first part of this thesis, we have developed a malaria transmission metamodel based on the susceptible-infected-resistant compartmental modeling framework (SIR) and taking into consideration human mobility flows between different villages in the Center of Senegal. Geographically targeted intervention strategies had been shown to be effective in reducing the incidence of malaria both within and outside of intervention areas. However, combined interventions targeting both vector and host, coordinated on a large scale are needed in regions and countries aiming to achieve malaria elimination in the short/medium term.In the second part we have evaluated different methods of estimating human mobility in the absence of real data. These methods included spatio-temporal traceability of mobile phones, mathematical models of gravity and radiation. The transport of the pathogen through the geographical space via the mobility of an infected subject is a major determinant of the spread of an epidemic. We introduced the impedance model that minimized the mean square error on mobility estimates, especially in contexts where population sets are characterized by their heterogeneous sizes.Finally, we have expanded the framework of assumptions underlying the calibration of the gravity models of human mobility. The hypothesis of a zero inflated distribution provided a better fit and a better predictability, compared to the classical approach not assuming an excess of zeros: Poisson, Quasipoisson.