Thesis advisor: H. Michael Mann / Thesis (PhD) — Boston College, 1974. / Submitted to: Boston College. Graduate School of Arts and Sciences. / Discipline: Economics.

No description available.

No description available.

No description available.

The empirical evidence concerning the associations between diet, particularly fat intake, and obesity is inconclusive. The aim of the present study was to investigate cross-sectional associations between general and central adiposity, and dietary and other socio-demographic and behavioural factors influencing energy balance. Study subjects were 961 women and 596 men aged 30-75 years who participated in the Greek segment of the European Prospective Investigation into Cancer and Nutrition (EPIC) Study. General obesity was evaluated by Body Mass Index (BMI) and central obesity by Waist to Hip ratio (WHR) and Waist Circumference (WC). Dietary intake was estimated through a validated semi-quantitative Food Frequency Questionnaire (FFQ) obtained at baseline. Time weighted occupational and leisure activities, as well as socio-demographic and behavioural data were assessed through a life-style questionnaire. The methodological issues related to under-reporting and adjustment for energy intake have been considered in depth. Obesity indices (BMI, WHR, WC) were initially regressed, separately for men and women, on energy intake and energy expenditure, after adjusting for the confounding effects of age, socio-economic status and smoking habits and controlling for dietary under-reporting. Results indicated that increasing physical activity is less effective than decreasing energy intake in reducing BMI. WHR and WC were not affected by energy intake, whilst energy expenditure reduced WHR and WC independently of BMI. Obesity indices (BMI, WHR, WC) were subsequently regressed on nutrient intake after controlling for the confounding effects of energy intake, energy expenditure, age, socio- economic status and smoking habits, including and excluding under-reporters of energy intake. Among women, but not men, the nutrient more strongly positively associated to BMI was protein and to a lesser extent mono-unsaturated and total fat intake. WHR and WC do not seem to be differentially affected by energy equivalent amounts of energy generating nutrients.

Abstract Population size and trends of large carnivores are difficult to determine, but are often needed to inform conservation actions. Direct counts maintained over long time periods are extremely difficult to achieve. Indices of population sizes can be used to estimate large carnivore abundances, but are often case-, species- and sitespecific. Here, we test the general applicability of track-based indices to estimate large carnivore abundance. We surveyed 15 306.4km of roads associated with 339 transects across a wide geographical scale, large range of densities and variable substrates for tracks of African large carnivores. A combined model for all carnivore species on sandy soils serves as a robust approach to predict large carnivore densities. Thus, indices based on track counts can provide useful estimates of carnivore abundance. We found consistent relationships between track densities and the actual carnivore densities, having taken account of substrate.

Dissertação (mestrado) - Universidade Federal de Santa Catarina, Centro Sócio-Econômico, Programa de Pós-Graduação em Contabilidade, Florianópolis, 2010 / Made available in DSpace on 2012-10-25T03:09:18Z (GMT). No. of bitstreams: 1 278342.pdf: 424678 bytes, checksum: f8588acad2889a08dd00023206f085aa (MD5) / Diante das novas práticas contábeis referentes aos Ativos Intangíveis (AIs), este estudo tem por objetivo analisar as informações evidenciadas sobre os AIs nas demonstrações contábeis das empresas brasileiras pertencentes ao Ibovespa nos anos de 2006, 2007 e 2008. A amostra deste estudo é composta por 47 companhias não financeiras, com maior volume de negociação na BM&FBOVESPA. A pesquisa caracteriza-se como teórico-empírica de caráter descritivo com abordagem qualitativa-quantitativa por meio da análise documental das demonstrações contábeis das organizações pesquisadas. Os resultados demonstram que, no ano de 2006, 35 empresas divulgaram em seus demonstrativos financeiros a norma da CVM ou do IBRACON, que trata da criação do subgrupo intangível. Em 2007, 29 companhias apresentaram informações sobre a criação do subgrupo intangível de acordo com a Lei n. 11.638/07. As práticas contábeis mais específicas sobre os AIs, provenientes do CPC, foram evidenciadas por 34 organizações em 2008. Em 2006, 27 empresas apresentaram os AIs de modo segregado nas demonstrações contábeis, seguidas por 33 em 2007, e 47 em 2008. Os tipos de AIs mais citados são o Ágio e o Software, seguidos pelo Direito Contratual. Os itens comuns às normas internacionais mais divulgados são o valor contábil, o método de amortização e a vida útil. Os AIs mais representativos são o Ágio, Direito Contratual, e o Software. Verifica-se que há uma correlação positiva entre os investimentos realizados em Ativos Intangíveis (Gi) e o Grau de Intangibilidade (GI) das empresas. Os modelos de regressão investigados são significativos e apontam uma relação positiva entre os AIs e as variáveis financeiras propostas. Conclui-se que os AIs são recursos estratégicos para as organizações. O seu controle proporciona vantagens competitivas, diferenciação no mercado e benefícios econômicos futuros acima da média em relação aos concorrentes. Por isso, a evidenciação contábil é fundamental para a compreensão das atividades organizacionais, bem como o desempenho econômico-financeiro das entidades.

A method of calculating the RP-HPLC retention indices, based on the alkylarylketone scale, has been developed. The retention indices are calculated from the molecular structure of a compound as the sum of the parent contribution, the parent index, the substituent contributions, the substituent indices, and terms to account for the interactions between substituents, the interaction indices.

Dans le cadre du contrôle qualité de décors verriers par analyse d’image, nous proposons une méthodologie originale de classification supervisée de textures couleur pour identifier les défauts d’aspect présents sur les décors. Cette méthodologie consiste à construire un espace d’attributs de texture couleur discriminant de dimension réduite lors d’un apprentissage hors ligne afin d’y représenter les textures à classer en ligne. Afin de satisfaire aux contraintes exigées par une application industrielle en termes de qualité de résultats et de temps de calcul, l’originalité de notre approche consiste à sélectionner automatiquement un nombre réduit d’attributs qui, d’une part sont évalués à partir d’images codées dans plusieurs espaces couleur exploitant des propriétés différentes et d’autre part, tiennent compte des relations spatiales intra et inter-composantes existant au sein de ces espaces. Nous montrons alors que les indices d’Haralick extraits des matrices de co-occurrences chromatiques sont des attributs répondant à ces objectifs lorsque le nombre de couleurs de l’image est réduit grâce à une sous-quantification des composantes couleur et qu’un voisinage isotropique est utilisé. L’approche proposée est d’abord appliquée à trois bases d’images de textures couleur de référence afin de montrer l’apport de l’approche multi-espaces couleur et le bénéfice que présente la sélection d’attributs, avant d’être appliquée au contrôle qualité des décors verriers. Pour répondre au problème de sous-représentativité des prototypes lié à cette application, nous introduisons une approche originale basée sur la génération d’images de synthèse présentant les défauts à détecter. / In the framework of decorated glasses quality control by image analysis, we propose an original methodology of supervised color texture classification in order to identify aspect flaws on patterns. This methodology consists in determining a low dimensional discriminating color texture feature space during an off-line learning stage in order to perform an on-line texture classification in this selected feature space. To satisfy constraints required by industrial applications about processing time and quality of texture classification, the originality of our approach consists in automatically selecting a reduced number of features which, on one hand are evaluated from images coded in several color spaces with different properties and on the other hand, which take into account the spatial relationships within and between the color components of each space. Then, we show that Haralick features extracted from color cooccurrence matrices answer to these goals when the number of colors in the image is reduced thanks to a quantization of color components and when an isotropic neighborhood is used. The proposed approach is firstly applied on three color texture benchmark databases in order to show the contribution of the multi color space approach and the advantage of feature selection. The method is then applied to decorated glasses quality control. In order to answer to the lack of prototypes, we introduce an original approach based on the generation of synthetic images where aspect flaws are present.

Issues related to global warming and weather changes have forced people start to pay attention to energy saving. We expect that Smart Home Energy Management (SHEM) would be an important development over the next decade. In some environments cost is important, in other environments living quality is important and in other environments a tradeoff between cost and living quality is important. SHEM means being able to manage electrical loads so as to meet different purposes in homes. In this thesis, we develop a SHEM to curtail some electrical loads at peak time to meet predefined circuit level demand limits while minimizing the effect on users¡¦ living quality. The core of our SHEM is an electrical control loop which is developed based on heuristic modifications through lots of case studies and trials. To this end, we study several utilization characteristics of household loads including air conditioning, water heaters clothes dryers, and electric vehicles and model their behaviors through computer simulations. Finally, we implement the whole ideal of our SHEM in LabVIEW (Laboratory Virtual Instrument Engineering Workbench). Several simulations are conducted to verify the robustness and efficiency of our SHEM. keyword : Quick Charge,Load Priority,Convience Preference,Severity Indices,Duration Indices.