As dificuldades de aprendizagem sobre estatística e suas aplicações simplesmente em razão da matemática envolvida, principalmente, pelos alunos dos cursos de humanas, fazem dela um desafio para o professor que a ministra e também para o aluno que aprende. Assim, esta tese busca mostrar o quanto à elaboração e o desenvolvimento de uma ferramenta de avaliação tipo teste em um ambiente virtual pode contribuir como apoio ao ensino presencial do estudo de ferramentas estatísticas multivariadas para os alunos de graduação em Administração da FEARP/USP, matriculados na disciplina Estatística Aplicada à Administração II. Este trabalho faz parte do projeto virtual denominado LaViE, que é um ambiente virtual de ensino-aprendizagem de estatística, e que está fundamentado em três dimensões: Pedagogia Virtual, Tecnologia da Comunicação e Processo de Validação. Para a criação dessa ferramenta de avaliação tipo teste online foi necessário, primeiramente, a elaboração de um protocolo para desenvolvimento e implementação desse sistema no LaViE. A metodologia, então, foi baseada em três etapas: 1) elaboração do protocolo fundamentado na pesquisa exploratória de pedagogia virtual com base na teoria sobre processo de ensino-aprendizagem para o desenvolvimento de um conjunto de regras e passos que balizasse a criação de questões testes on-line com níveis diferentes de ?adaptação? pelo aluno para cada módulo apresentado na disciplina em questão. Neste caso, foram criados três níveis de adaptação: básico (I), intermediário (II) e avançado (III); 2) implementação dessa ferramenta de avaliação no LaViE (teste seu conhecimento), com o desenvolvimento das questões baseadas nos assuntos sobre análise estatística multivariada e; 3) avaliação quantitativa da opinião dos alunos (usuários) quanto à usabilidade do sistema ?teste seu conhecimento? desenvolvido. Esses dados foram coletados quando essa disciplina foi ministrada, porém em dois períodos diferentes; no segundo semestre de 2005 como projeto-piloto, e no segundo semestre de 2006. Dessa forma, pôde ser feita uma análise comparativa da opinião dos alunos sobre o sistema em dois momentos diferentes. Para esse levantamento foram utilizados dois questionários aplicados diretamente aos alunos em sala de aula momentos antes da avaliação presencial da disciplina e um outro logo após essa avaliação presencial. Essas três etapas finalizadas deram condição de existência de uma ferramenta que faz parte de uma das três dimensões citadas do LaViE ? Processo de Validação. / The mathematical knowledge involved in the process of learning statistics and its applications, brings about difficulties to the Humanities department students, thus making it a challenge not only to those who teach it but also to the ones who learn it. Therefore, this work aims at showing the contribution the elaboration and development of a tool such as an online evaluation test can give, as a support to the presencial teaching of the study of multivariate statistical resources, to the FEARP/USP Business Management undergraduate students, registered in Applied Statistics to Business Management II. This study is part of a virtual project named LaViE which is a virtual environment of teaching-learning of Statistics based on three dimensions: Virtual Pedagogy, Communication Technology and Validation Process. For the development and implementation of the online evaluation test in the LaViE project, a protocol was elaborated. The methodology involved is based on three phases: 1) protocol production built on the virtual pedagogy exploratory research founded on the theory concerning the teaching-learning process in order to develop a set of rules and measures which support the creation of online test problems presenting different levels of adaptation for each module of the mentioned discipline. Thus, three levels of adaptation were created: basic (I), intermediate (II), and advanced (III); 2) implementation of such evaluation tool in the LaViE project (test your knowledge) by developing questions based on topics about multivariate statistical analysis; 3) quantitative evaluation of the users? opinion as to the usability of the ?test your knowledge? system. Data were collected in two distinct occasions: second semester of 2005, as a pilot project, and second semester of 2006, what enabled a comparative analysis of the system by the users at two different moments. This survey was conducted by means of two questionnaires completed in class by the students, being the first before the presencial discipline evaluation, and the second immediately after it. At the completion of the three phases previously mentioned, the Validation Process was obtained.

Existe uma polêmica sobre qual deve ser o papel social das empresas. De um lado, considera-se que a responsabilidade de seus gestores é maximizar os resultados em favor de seus acionistas e proprietários e, de outro, que a ação social empreendida pelas empresas traz benefícios à sociedade e também a seus stakeholders. Além disso, alguns autores admitem, inclusive, que não basta seguir as leis e pagar os impostos, mas também ter atitudes éticas e filantrópicas. Essa dissertação é um estudo empírico, de caráter exploratório, que objetiva identificar a existência de relações significativas entre o desempenho financeiro das empresas e o seu nível de investimento em ações de responsabilidade social. A pesquisa envolveu a análise de uma amostra de 897 balanços sociais publicados pelo IBASE no período de 2000 a 2004, por 298 empresas de diversos portes, nacionais e multinacionais, que atuavam no Brasil. Estudos científicos que buscam identificar relacionamentos entre ação social empresarial e seu impacto sobre o desempenho financeiro têm procurado responder, mas ainda não encontraram uma resposta definitiva para essa questão. Algumas dessas pesquisas foram abordadas na revisão bibliográfica demonstrando que o tema ainda carece de mais investigação. O presente estudo avaliou esse relacionamento sob várias condições: influência da situação financeira de lucro ou prejuízo das empresas, setor de atuação, seu porte e a conjuntura econômica e social em que elas operaram, refletida no ano de publicação de seu balanço social. O resultado da análise encontrou significância estatística que confirmou algumas das pesquisas científicas publicadas e também apresentou resultados que responderam questões em aberto de outras pesquisas sobre o mesmo tema. Portanto, apesar da dificuldade natural de percepção integrada dos resultados dos 68 modelos matemáticos envolvidos no estudo, as conclusões são uma oportuna contribuição científica ao tema. / There is a polemic about which should be the social role of companies. At one side, it is considered that the managers\' responsibility is to maximize the results for stockholders and owners, and at the other, the social action undertaken by companies to bring benefits to the society and also to the stakeholders. Besides, some authors admit that is not enough to abide to the laws and to pay taxes, but the firm must also have ethical and philanthropical attitudes. This dissertation is an empiric study, of exploratory nature, aiming at identifying the existence of significant relationships between financial performance of the companies and their investment level in social responsibility. The research involved the analysis of a sample of 897 social accounts published by IBASE in the period of 2000 to 2004, by 298 companies of several sizes, owned by national and multinational capital and operating in Brazil. Scientific studies that seek to identify relationships between managerial social action and their impact on the financial performance have searched to answer this question, but they did not find a definitive answer yet. Some of these researches were mentioned in the bibliographic revision demonstrating that the theme still needs more investigation. The present study evaluated this relationship under several conditions: companies having profit or loss, the business sector where they act, the firm size and the economic and social conjuncture in which they operated, reflected in the year of publication of their social accounting. The result of this analysis found statistical significance that confirmed some of the published scientific researches and also presented original results that answered open matters of other researches about the subject. Therefore, although the natural difficulty of integrated perception of the results of 68 mathematical models involved in the study, the conclusions are an opportune scientific contribution to the theme.

Heat of hydration is a property of Portland cement and a direct result of the chemical reaction between cement and water. The amount of heat released is dependent upon the cement mineralogical composition, curing temperature, water to cement ratio, and cement fineness. High temperature resulting from heat of hydration (thereon referred to as HOH) of cement can affect the hydration process, and consequently the kinetics of development of the mechanical properties of concrete. One of the main reasons triggering the interest in HOH of cement is its implication in thermal cracking of concrete. The high temperature gradient between the inner core and the outer surface of a concrete element is known to result in large tensile stresses that may exceed tensile strength, thus leading to early–age thermal cracking in mass concrete. This dissertation initially addresses accurately predicting the heat of HOH of Portland cement at seven days based on the heat flow data collected from isothermal calorimetry for a time interval of 0-84 h. This approach drastically reduces the time required to identify the seven day HOH of Portland cement. The second part of this study focuses on cement fineness and its critical role on the heat generated by Portland cement during hydration. Using a matrix of four commercially available Portland cements, representing a wide range of mineralogical composition, and subjecting each of the as-received cements to several grinding increments, a linear relationship was established between cement fineness and heat of hydration. The effect of cement fineness and mineralogical composition on HOH of Portland cement was then related through a mathematical expression to predict the HOH of Portland cement based on its mineralogical composition and fineness. Three expressions were proposed for the 1, 3 and 7 day HOH. The findings indicate that the equations developed, based on cement main phase composition and fineness, can be used to identify cements with high heat of HOH that may cause thermal cracking in mass concrete elements. Also, the equations can be used to correlate the HOH with the other properties of Portland cement for quality control and prediction of chemical and physical properties of manufactured Portland cement and concrete. Restrained shrinkage experiments results on mortar specimens prepared with cements of variable phase composition and fineness indicate that interaction of C3A and sulfate source is the prime phenomenon followed by cement fineness as the second main factor influencing concrete cracking. In order to minimize this effect, the third part of this study focused on studying alternatives that can lower the heat generated by concrete on hydration through the incorporation of nanomaterials; namely, graphene nanoparticles. The results indicate that incorporation of graphene a as replacement for Portland cement improves thermal diffusivity and electrical conductivity of the cement paste. Consequently, the use of graphene can trigger improvement of the thermal conductivity of concrete elements thus reducing the cracking potential of concrete. Measurements of HOH of graphene-cement paste, at w/c=0.5, using isothermal conduction calorimetry, indicate that incorporation of graphene up to 10% increases the length of the induction period while reduces the magnitude of the alite main hydration peak due to the filler effect. Furthermore, increasing the w/c ratio from 0.5 to 0.6 and graphene content from 1 % to 10% (as a partial replacement of cement) increases the 7 day HOH of Portland cement by 50 J/g. Isothermal conduction calorimetry heat flow curves show that incorporation of graphene particles up to 10% does not have significant effect on interaction of aluminates and sulfates sources since the time of occurrence of the C3A sulfate depletion peak is not affected by graphene substitution up to 10%. Full factorial statistical design and analysis conducted on compressive strength data of mortar specimens prepared at two w/c ratios, using cements of different finenesses and graphene content indicates that the quantity of graphene and the physical interaction due to variable w/c, graphene and cement fineness, have the smallest P-value among all the samples, representing the most significant impact on compressive strength of mortar samples. It appears that in graphene cement paste composites, addition of 1% graphene results in 21% reduction of Young’s modulus. Increasing the graphene content from 1% to 5% and/or 10% does not show significant effect on Young’s modulus. Similar trends can be observed in the hardness of graphene cement paste samples. In conclusion, partial replacement of Portland cement with graphene nanoparticles in concrete mixtures is a good alternative to lower the cracking potential in mass concrete elements.

This essay covers a designed experiment on paperboard where the quality under study is tear strength alongside and across. The objective is to examine what consequences the loss of balance in a designed experiment has on the explanatory power of the proposed empirical model. As did happen, the trial plan didn’t go as planned when the first run caused a disruption of the paperboard in the machine. Decision from the company was to raise the low level of one of the design factors to prevent this from happening again. The consequence of this is an alteration of the design during ongoing experimentation. This in turn affects what analysis approaches are appropriate for the problem. Three different approaches for analyzing the data are presented, each with different propositions on how to deal with the complication that occurred. The answer to the research question is that the ability of the empirical model to discover significant effects is moderately weakened by the loss of one run (out of eight total). The price payed for retrieving less information from the experiment is that the empirical model, for tear strength across, doesn’t deem the effects significant at the same level as for the candidate model with eight runs. Instead of concluding that the main effect of  and the interaction effect  is significant at the 2%- and 4%-level, respectively, we must now settle with deeming them significant at the 6%- and 7%-level.

Background: The development of the O2O market in China has been increasingly growing and will remain a steady growth in 2017 (iiMedia Research, 2016), but it is still an ambiguous phenomenon in the academic world and lack of relevant research, especially the influential factors to customer experience in the O2O business model. Research question: What factors affect customer experience in China in the context of O2O business? Purpose: The primary objective of this paper is to examine the influential factors of the customer experience in the context of O2O context, especially to the Chinese customers in the online travel industry. The purpose is to contribute to theoretical development on this topic. Method: The research is a quantitative study with deductive reasoning which was carried out by the survey. The data collection was conducted through online questionnaires. Conclusion: The findings of this study show that all five variables (reference group influence, perceived ease-of-use, sensory experience, brand association and spatial accessibility) are positively influential to Chinese customer experience in the context of O2O business. Perceived ease-of-use and Sensory experience are the most dominant influential factors followed by the Reference group influence and Spatial accessibility. The Brand association impacts customer O2O experience least.

Typically, a Parkinson’s disease (PD) patient would display instances of tremor and bradykinesia (slowness of movement) at an early stage of the disease and later develop gait disturbances and postural instability. So, it is important to measure the tremor occurrences in subjects to detect the onset of PD. Also, it is equally essential to monitor the gait impairments that the patient displays, as the order at which the PD symptoms appear in subjects vary from one to another. The primary goal of this thesis is to develop a monitoring system for PD patients using wearable sensors. To achieve that objective, our work focused first on identifying the most significant features that would best distinguish between PD and normal healthy subjects. Here, the various gait and tremor features were extracted from the raw data collected from the wearable sensors and further analyzed using statistical analysis and pattern classification techniques to pick the most significant features. In statistical analysis, the analysis of variance (ANOVA) test was conducted to differentiate the subjects based on the values of the mean. Further, pattern classification was carried out using the Linear Discriminant Analysis (LDA) algorithm. The analysis of our results shows that the features of heel force, step distance, stance and swing phases contributed more significantly to achieving a better classification between a PD and a normal subject, in comparison with other features. Moreover, the tremor analysis based on the frequency-domain characteristics of the signal including amplitude, power distribution, frequency dispersion, and median frequency was carried out to identify PD tremor from different types of artifacts.

>Magister Scientiae - MSc / Groundwater is one of the most valuable natural resources on earth and it forms an important part of the total water resources of South Africa. For this reason, this resource should be monitored and controlled on a regular basis. The study was conducted in Newcastle, in the KwaZulu-Natal province of South Africa. The aim of this study was to assess groundwater quality geochemically and determine its suitability for domestic and irrigation purposes. For the purpose of this study 31 samples were collected from 31 boreholes in and around the town of Newcastle. The samples were analysed for Magnesium (Mg2+), Calcium (Ca2+), Sodium (Na+), Potassium (K+), Chloride (Cl-), Sulfate (SO42-), Bicarbonate (HCO3-), Nitrate (NO3-), Fluoride (F-) (pH, TDS and Ec. The SAQWG (DWAF, 1996) and the WHO (2011) water standards were used as the basis of evaluating the suitability of groundwater for drinking purposes. For irrigation, Electrical Conductivity (EC), Sodium Percent (Na %), Sodium Adsorption Ratio (SAR), Residual Sodium Carbonate (RSC), Kelly’s Ratio (KR), Magnesium Ratio (MR) and Permeability Index (PI) were used to evaluate suitability. Classical hydro-chemical methods together with multivariate statistical methods were used to further understand the composition controlling processes. Lastly, the spatial distribution of the results was presented using ArcGIS. The results showed that the groundwater is alkaline in nature and that most of the samples are within the permissible range of both SAWQG (DWAF, 1996) and WHO (2011). Few samples showed concentration of Na+, Cl-, SO42- , F- and TDS above the guideline value as per WHO (2011) standards. The order of abundance of major ions in the groundwater, based on their mean values is as follows: Na+>Ca2+>Mg2+>K+ and HCO3->SO42- >Cl->NO3-. Classical hydro-chemical methods revealed four hydro-chemical facies in the study area, which are Ca-Mg-HCO3, Na-HCO3, Ca-Na-SO4-HCO3 and Na-Cl. The major ion chemistry analysis revealed that the main composition controlling processes in the study area is rock-water interaction. It further revealed that the ionic concentration is due to silicate weathering, carbonate weathering, cation exchange, gypsum dissolution and halite dissolution. Factor analysis indicated three factors, which explained 79.71 % of the total variance in the water quality data. The first factor which accounted for the highest variance in the data was the Alkalinity factor, followed by the Hardness factor and the Anthropogenic factor which accounted for the least variance. The cluster analysis revealed five clusters and discriminant analysis showed that Na+, TH, HCO3- and SO42- discriminate these clusters by 96.8%. In conclusion, the study revealed that the groundwater in most of the boreholes in the study area is generally suitable for drinking and irrigation. This is with exception to boreholes 13 and 31 which showed concentrations higher than the permitted level by WHO (2011) standards of TDS, Na+, Cl-, SO42- and F-. Similarly, 45.16% the groundwater samples showed that the groundwater has high sodium hazard potential which makes water from these boreholes unsuitable for irrigation purposes without proper treatment. accounted for the least variance. The cluster analysis revealed five clusters and discriminant analysis showed that Na+, TH, HCO3- and SO42- discriminate these clusters by 96.8%. accounted for the least variance. The cluster analysis revealed five clusters and discriminant analysis showed that Na+, TH, HCO3- and SO42- discriminate these clusters by 96.8%.In conclusion, the study revealed that the groundwater in most of the boreholes in the study area is generally suitable for drinking and irrigation. This is with exception to boreholes 13 and 31 which showed concentrations higher than the permitted level by WHO (2011) standards of TDS, Na+, Cl-, SO42- and F-. Similarly, 45.16% the groundwater samples showed that the groundwater has high sodium hazard potential which makes water from these boreholes unsuitable for irrigation purposes without proper treatment.

In big data systems, the infrastructure is such that large amounts of data are hosted away from the users. Information security is a major challenge in such systems. From the customer’s perspective, one of the big risks in adopting big data systems is in trusting the service provider who designs and owns the infrastructure, with data security and privacy. However, big data frameworks typically focus on performance and the opportunity for including enhanced security measures is limited. In this dissertation, the problem of mitigating insider attacks is extensively investigated and several static and dynamic run-time techniques are developed. The proposed techniques are targeted at big data systems but applicable to any data system in general. First, a framework is developed to host the proposed security techniques and integrate with the underlying distributed computing environment. We endorse the idea of deploying this framework on special purpose hardware and a basic model of the software architecture for such security coprocessors is presented. Then, a set of compile-time and run-time techniques are proposed to protect user data from the perpetrators. These techniques target detection of insider attacks that exploit data and infrastructure. The compile-time intrusion detection techniques analyze the control flow by disassembling program binaries while the run-time techniques analyze the memory access patterns of processes running on the system. The proposed techniques have been implemented as prototypes and extensively tested using big data applications. Experiments were conducted on big data frameworks such as Hadoop and Spark using cloud-based services. Experimental results indicate that the proposed techniques successfully detect insider attacks in the context of data loss, data degradation, data exposure and infrastructure degradation.

L’analyse statistique de la forme de structures anatomiques est un enjeu essentiel pour de nombreuses applications: modélisation de la variabilité normale et pathologique, prédiction de paramètres cliniques et biologiques à partir de données anatomiques… Ces dernières années ont vu l’émergence de grandes bases de données en neuroimagerie, offrant une puissance statistique considérablement accrue. Cette thèse a pour thème l’étude statistique de la variabilité anatomique de l’hippocampe sur de grandes populations. Après un état de l’art, la première partie de la thèse porte sur l’étude d’une variante anatomique appelée inversion incomplète de l’hippocampe (IHI). Pour ce faire, nous avons mis au point une échelle de cotation de ces IHI. Elle a été ensuite appliquée à 2000 sujets sains et jeunes de la base de données IMAGEN. Les résultats ont permis de montrer que les IHI sont fréquentes chez les sujets sains, avec une prédominance à gauche. C’est la première fois que les IHI sont étudiés sur une telle population. La seconde partie de la thèse porte sur la mise au point d’une méthode d’analyse statistique de formes utilisant les grandes déformations difféomorphiques et les courants mathématiques, utilisable pour l’analyse de grandes populations. Nous avons en particulier introduit une nouvelle approche rapide pour construire des prototypes anatomiques. Cette approche a été validée sur 1000 sujets sains jeunes de la base de données IMAGEN et 294 sujets (sujets sains âgés et patients atteints de maladie d’Alzheimer) de la base ADNI. Les résultats montrent que la méthode permet de représenter la variabilité anatomique de l’hippocampe avec un nombre raisonnable de dimensions. / Statistical shape analysis of anatomical structures is a key challenge for many applications: modeling of the normal and pathological variability, prediction of clinical or biological parameters from anatomical data... Recent years have seen the emergence of large databases in neuro-imaging, potentially increasing the statistical power of new studies. This thesis is about the statistical analysis of the anatomical variability of hippocampi in large populations. After a state of the art, the first part of the thesis focuses on the study of an anatomical variant named Incomplete Hippocampal Inversion (IHI). We developed a new robust scale for IHI assessments. We then applied the evaluation to 2000 young healthy subjects from the European database IMAGEN. Results show that IHI are frequent on healthy population, with a left side predominance. This is the first time that IHI are studied on a large database composed of healthy subjects. The second part of this thesis develops a method for the statistical analysis of shapes based on Large Diffeomorphic Deformations Metric Mapping and mathematical currents, applicable for the analysis of large datasets. In particular we have introduced a new fast approach for the estimation of anatomical templates. This approach has been validated on 1000 young healthy subjects of the IMAGEN database and on 294 subjects from the ADNI database (healthy aging subjects and patients with Alzheimer disease). Results show that the method allows the modeling of the anatomical variability of hippocampi with a reasonable number of dimensions.

This dissertation proposes to combine magnetic resonance imaging (MRI), positron emission tomography (PET) and a neuropsychological test, Mini-Mental State Examination (MMSE), as input to a multidimensional space for the classification of Alzheimer’s disease (AD) and it’s prodromal stages including amnestic MCI (aMCI) and non-amnestic MCI (naMCI). An assessment is provided on the effect of different MRI normalization techniques on the prediction of AD. Statistically significant variables selected for each combination model were used to construct the classification space using support vector machines. To combine MRI and PET, orthogonal partial least squares to latent structures is used as a multivariate analysis to discriminate between AD, early and late MCI (EMCI and LMCI) from cognitively normal (CN)s. In addition, this dissertation proposes a new effective mean indicator (EMI) method for distinguishing stages of AD from CN. EMI utilizes the mean of specific top-ranked measures, determined by incremental error analysis, to achieve optimal separation of AD and CN. For AD vs. CN, the two most discriminative volumetric variables (right hippocampus and left inferior lateral ventricle), when combined with MMSE scores, provided an average accuracy of 92.4% (sensitivity: 84.0%; specificity: 96.1%). MMSE scores were found to improve classification accuracy by 8.2% and 12% for aMCI vs. CN and naMCI vs. CN, respectively. Brain atrophy was almost evenly seen on both sides of the brain for AD subjects, which was different from right side dominance for aMCI and left side dominance for naMCI. Findings suggest that subcortical volume need not be normalized, whereas cortical thickness should be normalized either by intracranial volume or the mean thickness. Furthermore, MRI and PET had comparable predictive power in separating AD from CN. For the EMCI prediction, cortical thickness was found to be the best predictor, even better than using all features together. Validation with an external test set demonstrated that best of feature-selected models for the LMCI group was able to classify 83% of the LMCI subjects. The EMI-based method achieved an accuracy of 92.7% using only MRI features. The performance of the EMI-based method along with its simplicity suggests great potential for its use in clinical trials.