Location analysis of city sections: socio-demographic segmentation and restaurant potentiality estimation : a case study of Lisbon

One of the objectives of this study is to perform classification of socio-demographic components for the level of city section in City of Lisbon. In order to accomplish suitable platform for the restaurant potentiality map, the socio-demographic components were selected to produce a map of spatial clusters in accordance to restaurant suitability. Consequently, the second objective is to obtain potentiality map in terms of underestimation and overestimation in number of restaurants. To the best of our knowledge there has not been found identical methodology for the estimation of restaurant potentiality. The results were achieved with combination of SOM (Self-Organized Map) which provides a segmentation map and GAM (Generalized Additive Model) with spatial component for restaurant potentiality. Final results indicate that the highest influence in restaurant potentiality is given to tourist sites, spatial autocorrelation in terms of neighboring restaurants (spatial component), and tax value, where lower importance is given to household with 1 or 2 members and employed population, respectively. In addition, an important conclusion is that the most attractive market sites have shown no change or moderate underestimation in terms of restaurants potentiality.

Adverse drug reaction classification with deep neural networks

We study the problem of detecting sentences describing adverse drug reactions (ADRs) and frame the problem as binary classification. We investigate different neural network (NN) architectures for ADR classification. In particular, we propose two new neural network models, Convolutional Recurrent Neural Network (CRNN) by concatenating convolutional neural networks with recurrent neural networks, and Convolutional Neural Network with Attention (CNNA) by adding attention weights into convolutional neural networks. We evaluate various NN architectures on a Twitter dataset containing informal language and an Adverse Drug Effects (ADE) dataset constructed by sampling from MEDLINE case reports. Experimental results show that all the NN architectures outperform the traditional maximum entropy classifiers trained from n-grams with different weighting strategies considerably on both datasets. On the Twitter dataset, all the NN architectures perform similarly. But on the ADE dataset, CNN performs better than other more complex CNN variants. Nevertheless, CNNA allows the visualisation of attention weights of words when making classification decisions and hence is more appropriate for the extraction of word subsequences describing ADRs.

Hilbert Space Valued Signal Plus Noise Models: Analysis of Structural Breaks under High Dimensionality and Temporal Dependence

This thesis is concerned with change point analysis for time series, i.e. with detection of structural breaks in time-ordered, random data. This long-standing research field regained popularity over the last few years and is still undergoing, as statistical analysis in general, a transformation to high-dimensional problems. We focus on the fundamental »change in the mean« problem and provide extensions of the classical non-parametric Darling-Erdős-type cumulative sum (CUSUM) testing and estimation theory within highdimensional Hilbert space settings. In the first part we contribute to (long run) principal component based testing methods for Hilbert space valued time series under a rather broad (abrupt, epidemic, gradual, multiple) change setting and under dependence. For the dependence structure we consider either traditional m-dependence assumptions or more recently developed m-approximability conditions which cover, e.g., MA, AR and ARCH models. We derive Gumbel and Brownian bridge type approximations of the distribution of the test statistic under the null hypothesis of no change and consistency conditions under the alternative. A new formulation of the test statistic using projections on subspaces allows us to simplify the standard proof techniques and to weaken common assumptions on the covariance structure. Furthermore, we propose to adjust the principal components by an implicit estimation of a (possible) change direction. This approach adds flexibility to projection based methods, weakens typical technical conditions and provides better consistency properties under the alternative. In the second part we contribute to estimation methods for common changes in the means of panels of Hilbert space valued time series. We analyze weighted CUSUM estimates within a recently proposed »high-dimensional low sample size (HDLSS)« framework, where the sample size is fixed but the number of panels increases. We derive sharp conditions on »pointwise asymptotic accuracy« or »uniform asymptotic accuracy« of those estimates in terms of the weighting function. Particularly, we prove that a covariance-based correction of Darling-Erdős-type CUSUM estimates is required to guarantee uniform asymptotic accuracy under moderate dependence conditions within panels and that these conditions are fulfilled, e.g., by any MA(1) time series. As a counterexample we show that for AR(1) time series, close to the non-stationary case, the dependence is too strong and uniform asymptotic accuracy cannot be ensured. Finally, we conduct simulations to demonstrate that our results are practically applicable and that our methodological suggestions are advantageous.

Evaluation of breeding strategies for polledness in dairy cattle using a newly developed simulation framework for quantitative and Mendelian traits

Background: Intensified selection of polled individuals has recently gained importance in predominantly horned dairy cattle breeds as an alternative to routine dehorning. The status quo of the current polled breeding pool of genetically-closely related artificial insemination sires with lower breeding values for performance traits raises questions regarding the effects of intensified selection based on this founder pool. Methods: We developed a stochastic simulation framework that combines the stochastic simulation software QMSim and a self-designed R program named QUALsim that acts as an external extension. Two traits were simulated in a dairy cattle population for 25 generations: one quantitative (QMSim) and one qualitative trait with Mendelian inheritance (i.e. polledness, QUALsim). The assignment scheme for qualitative trait genotypes initiated realistic initial breeding situations regarding allele frequencies, true breeding values for the quantitative trait and genetic relatedness. Intensified selection for polled cattle was achieved using an approach that weights estimated breeding values in the animal best linear unbiased prediction model for the quantitative trait depending on genotypes or phenotypes for the polled trait with a user-defined weighting factor. Results: Selection response for the polled trait was highest in the selection scheme based on genotypes. Selection based on phenotypes led to significantly lower allele frequencies for polled. The male selection path played a significantly greater role for a fast dissemination of polled alleles compared to female selection strategies. Fixation of the polled allele implies selection based on polled genotypes among males. In comparison to a base breeding scenario that does not take polledness into account, intensive selection for polled substantially reduced genetic gain for this quantitative trait after 25 generations. Reducing selection intensity for polled males while maintaining strong selection intensity among females, simultaneously decreased losses in genetic gain and achieved a final allele frequency of 0.93 for polled. Conclusions: A fast transition to a completely polled population through intensified selection for polled was in contradiction to the preservation of high genetic gain for the quantitative trait. Selection on male polled genotypes with moderate weighting, and selection on female polled phenotypes with high weighting, could be a suitable compromise regarding all important breeding aspects.

Iowa Seat Belt Use Survey 2013 Data Collection Methodology Report, September 26, 2013

In an effort to achieve greater consistency and comparability in state-wide seat belt use reporting, the National Highway Traffic Safety Administration (NHTSA) issued new requirements in 2011 for observing and reporting future seat belt use. The requirements included the involvement of a qualified statistician in the sampling and weighting portions of the process as well as a variety of operational details. The Iowa Governor’s Traffic Safety Bureau contracted with Iowa State University’s Survey & Behavioral Research Services (SBRS) in 2011 to develop the study design and data collection plan for the State of Iowa annual survey that would meet the new requirements of the NHTSA. A seat belt survey plan for Iowa was developed by SBRS with statistical expertise provided by Zhengyuan Zhu, Ph.D., Associate Professor of Statistics at Iowa State University and was approved by NHTSA on March 19, 2012.

Iowa Seat Belt Use Survey 2014 Data Collection Methodology Report, September 25, 2014

In an effort to achieve greater consistency and comparability in state-wide seat belt use reporting, the National Highway Traffic Safety Administration (NHTSA) issued new requirements in 2011 for observing and reporting future seat belt use. The requirements included the involvement of a qualified statistician in the sampling and weighting portions of the process as well as a variety of operational details. The Iowa Governor’s Traffic Safety Bureau contracted with Iowa State University’s Survey & Behavioral Research Services (SBRS) in 2011 to develop the study design and data collection plan for the State of Iowa annual survey that would meet the new requirements of the NHTSA. A seat belt survey plan for Iowa was developed by SBRS with statistical expertise provided by Zhengyuan Zhu, Ph.D., Associate Professor of Statistics at Iowa State University and was approved by NHTSA on March 19, 2012.

Iowa Seat Belt Use Survey 2015 Data Collection Methodology Report, September 9, 2015

In an effort to achieve greater consistency and comparability in state-wide seat belt use reporting, the National Highway Traffic Safety Administration (NHTSA) issued new requirements in 2011 for observing and reporting future seat belt use. The requirements included the involvement of a qualified statistician in the sampling and weighting portions of the process as well as a variety of operational details. The Iowa Governor’s Traffic Safety Bureau contracted with Iowa State University’s Survey & Behavioral Research Services (SBRS) in 2011 to develop the study design and data collection plan for the State of Iowa annual survey that would meet the new requirements of the NHTSA. A seat belt survey plan for Iowa was developed by SBRS with statistical expertise provided by Zhengyuan Zhu, Ph.D., Associate Professor of Statistics at Iowa State University and Director of the Center for Survey Statistics and Methodology. The plan was approved by NHTSA on March 19, 2012.

L'Intégration d'un Module d'Étude Internet Dans la Classe de Comptabilité: Perceptions des Élèves et les Résultats Académiques

Implemented in the context of Business Administration students enrolled in a college level three year technology program, this research investigated students’ perceptions and academic results concurrent with the implementation of an online web module designed to facilitate student self-study. The students involved in this research were enrolled in a program that, while offering a broad education in business disciplines, specialized in the field of accounting. As a result, students were enrolled in academically rigorous accounting courses in each of the six semesters of the program. The weighting of these accounting courses imposes a significant self-study component – typically matching or exceeding the time spent in class. In this context many of the students enrolled in the Business Administration Program have faced difficulties completing the self-study component of the course effectively as demonstrated in low homework completion rates, low homework grade averages and ultimately low success rates in the courses. In an attempt to address this situation this research studied the implementation of a web-based self-study module. Through this module students could access a number of learning tools that were designed to facilitate the self-study process under the premise that more effective self-study learning tools will help remove obstacles and provide more timely confirmation of learning during student self-study efforts. This research collected data from a single cohort of students drawn from the first three sequential accounting courses of the Business Administration Program. The web-based self-study module was implemented in the third of the three sequential accounting courses. The first two of these courses implemented a traditional manual self-study environment. Data collected from the three accounting courses included homework completion rates, homework, exam and final grades for the respective courses. In addition the web-study module allowed the automatic reporting of student usage of a number of specific online learning tools. To complement the academic data, students were surveyed to gain insight into their perceptions of the effectiveness of the web-based system. The research provided a number of interesting insights. First among these was a confirmation of the importance of the self-study process in the academic achievement of the learners. Regardless of the self-study environment, manual or web-enhanced, a significant positive correlation existed between the students’ self-study results, demonstrated in both homework completion rates and homework averages and the corresponding final grades. These results confirm the importance of self-study found generally in the prevailing academic literature regarding students enrolled in higher education. In addition, the web-enhanced learning environment implemented during the third accounting course coincided with significantly higher homework completion rates and corresponding homework averages: homework completion rates in particular increased from a combined average of 63% in the first two accounting courses to 93% in the web-enhanced context of the third accounting course. Moreover, the homework completion rates of the web-enhanced course were evenly distributed across the cohort of students. A quartile-based analysis was subsequently completed. Quartiles were constructed by ranking the students according to their combined average homework completion rates from the first two manual self-study courses, Accounting I and II. The quartile-based homework completion rates for the manual self-study courses Accounting I and II were subsequently compared to the results these same quartiles of students achieved in the web-based self-study within Accounting III. While the first two courses demonstrated significantly uneven homework completion rates across the quartiles ranging from 31% to 91% homework completion rates, the differences among the four quartiles within the web-enhanced module, with an average homework completion rate of 93%, were statistically insignificant. Congruent with the positive academic results observed in the third, web-enhanced course, through the corresponding survey, students expressed a strong attitude in favor of the online self-study environment. This research was designed to add to the existing research that studies the implementation of learning in an online setting. Specifically, the research was designed to explore a middle ground of online learning – a web-enhanced course – a context that supplements the classroom experience rather than replacing it. The web-enhanced accounting course demonstrated impressive favorable results, both academically and in terms of students' perception of the system; these results suggest that a web-enhanced environment can provide learning tools that facilitate the self-study process while providing a structured learning environment that can help developing learners reach their potential.

Optimisation de l’extraction de lumière de scintillation dans les matrices de détecteurs pixellisés pour des applications en tomographie d’émission par positrons (TEP) et en tomodensitométrie (TDM)

Résumé : Les performances de détecteurs à scintillation, composés d’un cristal scintillateur couplé à un photodétecteur, dépendent de façon critique de l’efficacité de la collecte et de l’extraction des photons de scintillation du cristal vers le capteur. Dans les systèmes d’imagerie hautement pixellisés (e.g. TEP, TDM), les scintillateurs doivent être arrangés en matrices compactes avec des facteurs de forme défavorables pour le transport des photons, au détriment des performances du détecteur. Le but du projet est d’optimiser les performances de ces détecteurs pixels par l'identification des sources de pertes de lumière liées aux caractéristiques spectrales, spatiales et angulaires des photons de scintillation incidents sur les faces des scintillateurs. De telles informations acquises par simulation Monte Carlo permettent une pondération adéquate pour l'évaluation de gains atteignables par des méthodes de structuration du scintillateur visant à une extraction de lumière améliorée vers le photodétecteur. Un plan factoriel a permis d'évaluer la magnitude de paramètres affectant la collecte de lumière, notamment l'absorption des matériaux adhésifs assurant l'intégrité matricielle des cristaux ainsi que la performance optique de réflecteurs, tous deux ayant un impact considérable sur le rendement lumineux. D'ailleurs, un réflecteur abondamment utilisé en raison de ses performances optiques exceptionnelles a été caractérisé dans des conditions davantage réalistes par rapport à une immersion dans l'air, où sa réflectivité est toujours rapportée. Une importante perte de réflectivité lorsqu'il est inséré au sein de matrices de scintillateurs a été mise en évidence par simulations puis confirmée expérimentalement. Ceci explique donc les hauts taux de diaphonie observés en plus d'ouvrir la voie à des méthodes d'assemblage en matrices limitant ou tirant profit, selon les applications, de cette transparence insoupçonnée.

Assessing Household Food Waste in Italy: A Methodology for Detecting Drivers and Quantities

It has been estimated that one third of edible food destined for human consumption is lost or wasted along the food supply chain globally. Much of the waste comes from Global North, where consumers are considered as the bigger contributors. Different studies tried to analyze and estimate the Household Food Waste (HFW), especially in UK and Northern Europe. The result is that accurate studies at national level exist only in UK, Finland and Norway while no such studies are available in Italy, except for survey- based researches. Though, there is a widespread awareness that such methods might be not able to estimate Food Waste. Results emerging from literature clearly suggest that survey estimate inferior amounts of Food Waste as a result, if compared to waste sorting and weighting analysis or to diary studies. The hypothesis that household food waste is under-estimated when gathered through questionnaires has been enquired into. First, a literature review of behavioral economics and heuristics has been proposed; then, a literature review of the sector listing the existing methodologies to gather national data on Household Food Waste has been illustrated. Finally, a pilot experiment to test a mixed methodology is proposed. While literature suggests that four specific cognitive biases might be able to affect the reliability of answers in questionnaires, results of the present experiment clearly indicate that there is a relevant difference between how much the individual thinks to waste and he/she actually does. The result is a mixed methodology based on questionnaire, diary and waste sorting, able to overcome the cons of each single method.