The classification task requires learning a decision boundary between classes by making use of training examples from each. A potential challenge for this task is the class imbalance problem, which occurs when there are many training instances available for a single class, the majority class, and few training instances for the other, the minority class [58]. In this case, it is no longer clear how to separate the majority class from something for which we have little to no knowledge. More worrying, often the minority class is the class of interest, e.g. for detecting abnormal conditions from streaming sensor data. The one-class classification (OCC) paradigm addresses this scenario by casting the task as learning a decision boundary around the majority class with no need for minority class instances [110]. OCC has been thoroughly investigated, e.g. [20, 60, 90, 110], and many one-class classifiers have been proposed. One approach for improving one-class classifier performance on static data sets is learning in the context of concepts: the majority class is broken down into its constituent sub-concepts and a classifier is induced over each [100]. Modern machine learning research, however, is concerned with data streams: where potentially infinite amounts of data arrive quickly and need to be processed as they arrive. In these cases it is not possible to store all of the instances in memory, nor is it practical to wait until “the end of the data stream” before learning. An example is network intrusion detection: detecting an attack on the computer network should occur as soon as practicable. Many one-class classifiers for data streams have been described in the literature, e.g. [33, 108], and it is worth investigating whether the approach of learning in the context of concepts can be successfully applied to the OCC task for data streams as well. This thesis identifies that the idea of breaking the majority class into subconcepts to simplify the OCC problem has been demonstrated for static data sets, [100], but has not been applied in data streams. The primary contribution to the literature made by this thesis is the identification of how the majority class’s sub-concept structure can be used to improve the classification performance of streaming one-class classifiers while mitigating the challenges posed by the data stream environment. Three frameworks are developed, each using this knowledge to a different degree. These are applied with a selection of streaming one-class classifiers to both synthetic and benchmark data streams with performance compared to that of the one-class classifier learning independently. These results are analyzed and it is shown that scenarios exist where knowledge of sub-concepts can be used to improve one-class classifier performance.

Many studies have interpreted the positive and negative consequences of China's One-Child Policy on Chinese women, but few have relied upon contribution from Chinese women and only children themselves. However, by valuing personal, lived experience, researchers may discourage the propagation of Western media stereotypes of what it means to be a young Chinese woman in urban China. The use of an intersectionality framework showcases how Chinese women's lived experiences must be framed more widely than a single aspect of their identity and that gender, educational experiences, family dynamic, and single child status (among other identity markers) form a more complex and holistic identity than any marker considered individually. This thesis argues that the intersectionality of lived experience makes it necessary to consider each individual's story as valuable and that commonalities and differences are both crucial to understanding how individual's lives intersect in the context of this complex birth policy

Nos travaux de thèse portent sur la mise au point de nouvelles méthodes de synthèse permettant l’accès de façon rapide et efficace à différents dérivés pyrazolo[1,5-a]pyrimidiniques diversement fonctionnalisés.La première partie de ce travail concerne l’étude et l’application de la condensation de la pyran-2-one avec les 5(3)-amino-3(5)-arylpyrazoles que ce soit par chauffage classique ou sous irradiation micro-onde. Par la suite, un nouveau procédé d’(hétéro)arylation pallado-catalysé direct régiosélectif a été mis à profit pour élaborer des pyrazolo[1,5-a]pyrimidines fonctionnalisées tant en position 3 que sur le méthyle situé sur le sommet 7 de la 5,7-diméthylpyrazolo[1,5-a]pyrimidine.Le dernier volet de ce mémoire a été consacré à la préparation des entités pyrazolo[1,5-a]pyrimidiniques substituées en position 7 par des motifs benzyliques et ce, via une procédure « one-pot », sous irradiation micro-onde, constituée d’une réaction d’arylation directe pallado-catalysée suivie d’une saponification-décarboxylation. / Our thesis works concern the development of new methods of synthesis allowing a rapid and effective access to variously functionalized pyrazolo [1,5-a] pyrimidines derivatives.First, we were explored the study and the application of the condensation of the pyran-2 one with 5(3)-amino-3 (5)-arylpyrazoles whether it is by classic heating or under microwaves. Indeed, we have described the first example of the direct and regioselective palladium-catalyzed (hetero)arylation of 5,7-dimethylpyrazolo[1,5-a]pyrimidine, the reaction may be inducted to occur at either an sp2 or an sp3 carbon atom.The last part of this report was dedicated to the preparation of some 7-substituted pyrazolo[1,5-a]pyrimidines via a one-pot two steps palladium-catalyzed direct CH-arylation followed by a saponification-decarboxylation.

abstract: Machine learning methodologies are widely used in almost all aspects of software engineering. An effective machine learning model requires large amounts of data to achieve high accuracy. The data used for classification is mostly labeled, which is difficult to obtain. The dataset requires both high costs and effort to accurately label the data into different classes. With abundance of data, it becomes necessary that all the data should be labeled for its proper utilization and this work focuses on reducing the labeling effort for large dataset. The thesis presents a comparison of different classifiers performance to test if small set of labeled data can be utilized to build accurate models for high prediction rate. The use of small dataset for classification is then extended to active machine learning methodology where, first a one class classifier will predict the outliers in the data and then the outlier samples are added to a training set for support vector machine classifier for labeling the unlabeled data. The labeling of dataset can be scaled up to avoid manual labeling and building more robust machine learning methodologies. / Dissertation/Thesis / Masters Thesis Engineering 2017

Submitted by Luciana Sebin (lusebin@ufscar.br) on 2016-10-11T12:21:35Z No. of bitstreams: 1 DissASA.pdf: 1261155 bytes, checksum: c70182e69e5460100b08d408d4eeb58c (MD5) / Approved for entry into archive by Marina Freitas (marinapf@ufscar.br) on 2016-10-20T13:40:04Z (GMT) No. of bitstreams: 1 DissASA.pdf: 1261155 bytes, checksum: c70182e69e5460100b08d408d4eeb58c (MD5) / Approved for entry into archive by Marina Freitas (marinapf@ufscar.br) on 2016-10-20T13:40:12Z (GMT) No. of bitstreams: 1 DissASA.pdf: 1261155 bytes, checksum: c70182e69e5460100b08d408d4eeb58c (MD5) / Made available in DSpace on 2016-10-20T13:40:18Z (GMT). No. of bitstreams: 1 DissASA.pdf: 1261155 bytes, checksum: c70182e69e5460100b08d408d4eeb58c (MD5) Previous issue date: 2016-01-29 / Não recebi financiamento / The perspective of an accelerated growth of the Internet of Things, has been increasing the concern on how to make the infrastructure of networks more robust and energetically efficient. In this sense, opportunistic and dynamic networks, such as delay and disruption tolerant networks (DTN) and vehicular ad-hoc networks (VANET), need to be compatible with extremely efficient message routing algorithms. Examples of routing algorithms for opportunistic networks include the Epidemic and the Spray-and-Wait, which have high message transmission cost. With the objective of obtaining a more efficient communication, this work presents a new message routing approach based on dynamic graphs, known as Time-Varying Graphs (TVG). In the literature, there are algorithms which calculate optimal paths within a TVG, for example, the path which allows the minimum travel time for a message. This work proposes a distributed routing algorithm, named Time-Varying Shortest Path (TVFP), which calculates the optimal path in a TVG that allows message delivery within the shortest possible time. The algorithm was implemented in the ONE simulator and experiments were conducted in order to comparatively evaluate its performance against other DTN algorithms. The results show that TVFP has better performance than the other algorithms, delivering messages in shorter times and causing a smaller overhead due to generation of redundant and intermediate messages. / A perspectiva de um crescimento acelerado da Internet das Coisas tem aumentado a preocupação em tornar as redes sem infraestrutura mais robustas e eficientes energeticamente. Nesse sentido, redes oportunistas e dinâmicas, tais como as redes tolerantes a atrasos e desconexões (DTN) e as redes veiculares ad-hoc (VANET), precisam ser compatíveis com algoritmos de roteamento de mensagens mais eficientes. Exemplos de algoritmos de roteamento em redes DTN incluem o Epidemic e o Spray-and-Wait, que possuem alto custo de transmissão de mensagens. Com o objetivo de se obter uma comunicação mais eficiente, este trabalho apresenta uma nova abordagem de roteamento de mensagens baseada no uso de grafos dinâmicos, conhecidos no inglês como Time-Varying Graphs (TVG). Na literatura, existem algoritmos que calculam caminhos ótimos em TVGs, como por exemplo, busca do caminho que permite o menor tempo de viagem. Este trabalho propõe a criação de um algoritmo distribuído de roteamento e encaminhamento de mensagens em redes DTN, denominado Time-Varying Fastest Path (TVFP), que calcula o caminho a partir do TVG para se entregar a mensagem ao destino no menor tempo possível, utilizando uma única instância da mensagem. O algoritmo foi implementado no simulador ONE e experimentos foram conduzidos para avaliar seu desempenho comparativamente com outros algorimos para DTNs. Os resultados mostram que o TVFP obtem melhor desempenho que os demais algoritmos, fazendo com que a mensagem chegue mais cedo ao destino, e causando uma menor sobrecarga de geração de mensagens redundantes e intermediárias.

This thesis includes multinominals with one{\crq}s variables. Aim is create collection exercises hereto subject. Collection is divided on chaps with given to problems. In every chapter is totality theory buckthorn examples and straddle examples. All choice example are exemplary processed. Behind every chapter reader can try out problems for examles that are supplementeds record

No description available.

No description available.

The purpose of this essay is to examine the relation between reproductive rights and overpopulation. The main research question is whether the increase of the population size can affect the moral aspects of the reproductive rights. Relying on, and building upon arguments already developed by a number of philosophers, I show that although some restrictions might be justified, these restrictions cannot completely rule out the basic right to reproduction. Based on this, I then proceed with an overview of the possible solutions. In the subsequent discussion, a one-child-per-couple philosophy stands out as the best option. This would be morally justified according to the initial statements, as well as the most effective alternative when it comes to initiating a significant population decline. The one-child-per-couple philosophy is then contested by several problems that could theoretically arise, but none of these problems are, according to my discussion, strong enough to dispute the positive aspects of the theory. The conclusion of this essay is that we, as a species, have a moral obligation to limit our numbers and that we ought to start acting accordingly.

The work presented focuses on bacterial zoonoses in northern Nigeria, and more specifically on brucellosis in the Kachia grazing reserve (KGR) - rangeland set-aside by the government to sedentarise Fulani pastoralists. The objectives of the study were to 1) undertake demographic and socioeconomic profiling of the KGR community; 2) review the evidence for brucellosis burden in Nigeria; 3) assess the suitability and performance of brucellosis diagnostic tests selected for use; 4) compare burden of brucellosis across different species (animal and human) and determine Brucella species present in KGR; 5) explore social or environmental factors which may promote or prevent brucellosis transmission; 6) make recommendations for brucellosis control in the KGR and Nigeria; 7) explore community perception of disease and determine household expenditure on animal health; 8) critically evaluate the system’s, integrated, disease cluster, ‘One Health’ approach applied in this study. Three surveys comprising animal (cattle, sheep and goat) and human sampling, administration of questionnaires, focus group discussions and key informant interviews were undertaken in March, June and October 2011. A population census was undertaken in June 2011. Comparison of 2010 government census data with June 2011 census data showed that a mass immigration event occurred in April-May 2011 as a result of post-election violence, with cattle and human populations increasing by 75%. Questionnaire and census data demonstrated the diversity and heterogeneity of the Fulani community in terms of wealth status (roughly corresponding to livestock assets), household size and composition and livelihood diversification strategies. While Fulani in grazing reserves were assumed to be sedentary, KGR households were found to practice wide-range dry and wet season transhumance. Cattle productivity parameters and herd dynamics were similar to those reported by other authors for the extensive pastoralist systems in the sub-humid zone. Herd increase over a one-year period was found to be low or negative for most households in this low input, low output system. Brucellosis epidemiology in the KGR involves B. abortus biovar 3a with low individual and moderate cattle herd prevalence and occasional spill-over into small ruminants. No human brucellosis was detected despite over 80% of the KGR population consuming raw milk and engaging in risky behaviours, raising questions about the potential lower virulence of the local biovar. Low infection rates in livestock, disease-reducing intuitive behaviours or immunity may also be at play. The RBT was found to perform well under field conditions, despite poor concordance when applied in different laboratories and under different conditions. Prospects for control/elimination of brucellosis in the KGR are poor, but low animal burden and absence of human disease render vaccination uneconomic. A review of the literature in Nigeria suggests that brucellosis burden is higher in intensive livestock production systems, which should be targeted first. A laissez-faire approach to brucellosis control in the nomadic pastoralist domain may appeal to policy-makers, as interventions in migratory populations are difficult. Brucellosis is perceived by the KGR community as the number three-priority disease, after trypanosomiasis and Fasciola gigantica/clostridial infection and this was reflected in household expenditure on chemotherapeutics and prophylaxis. Finally, the value of the One Health approach is the ability to see the whole picture, including disease impacts in the animal reservoir as well as the human population, without which erroneous epidemiological and economic conclusions may be drawn; for example, presence of brucellosis in the animal reservoir does not necessarily indicate presence of human disease. This work shows that moving from disciplinary silos to a more holistic or system’s approach spanning epidemiology, evaluation of diagnostic and control tools as well as socio-economic, cultural and institutional aspects can lead to more appropriate recommendations for disease control.