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...

Em nossa prática docente no ensino da Estatística para os cursos de engenharia temos constatado a dificuldade dos alunos no entendimento dos conceitos envolvidos nos métodos estatísticos, tendo como conseqüência a falta de motivação para a sua aprendizagem e, em geral, um elevado índice de reprovação. Passamos, então, a refletir sobre as causas dessa dificuldade e os meios de eliminá-las. Verificamos que o caráter problemático do ensino da Estatística decorre de uma equivocada visão da realidade, conseqüência da pouca familiaridade dos alunos com os fenômenos aleatórios que, embora estejam presentes em seu cotidiano, devido ao caráter excessivamente determinista dos currículos escolares, em geral, não são estudados no ensino fundamental e médio. Os objetivos do presente trabalho consistem em: (1) explicitar uma concepção da realidade em que o equilíbrio determinístico/aleatório seja restaurado; (2) repensar o ensino da Probabilidade e da Estatística nos diversos níveis tendo em vista tal equilíbrio e, a partir dele; (3) propor uma nova organização da disciplina Estatística nos cursos de graduação em Engenharia. Fomos buscar no pensamento filosófico e na evolução das idéias da ciência física a concepção predominante sobre os aspectos determinísticos e aleatórios dos fenômenos naturais...

Esta investigação verificou a aplicabilidade da Educação a Distância com vistas a instrumentar um público de acadêmicos, docentes e profissionais da área de saúde, para o uso do programa de banco de dados e estatística EPI INFO. Os objetivos da investigação foram planejar o conteúdo programático e implementar um curso online sobre construção e manuseio de banco de dados de estatística EPI INFO, verificar as facilidades e as dificuldades na implantação da Educação a Distância (EaD) e descrever a avaliação dos alunos relativa ao Curso no Ambiente Virtual de Aprendizagem. O Ambiente Virtual de Aprendizagem (AVA) utilizado para desenvolvimento do Curso foi CoL (Cursos Online). O conteúdo foi desenvolvido e organizado em quatro unidades: Unidade 1: Criação do Questionário; Unidade 2: Entrando com Dados; Unidade 3: Gerenciamento Básico de Dados na Análise; Unidade 4: Análise Intermediária: Estatísticas Básicas e Recuperação dos Dados. Foram inscritos 32 alunos; destes, 10 participaram efetivamente. Os alunos que completaram o Curso avaliaram de maneira satisfatória a utilização da Internet como ferramenta de aprendizado. O Curso Epi Info desenvolvido na modalidade a distância apresentou, nos itens avaliados...

O objetivo desta dissertação foi desenvolver um método para implementação e acompanhamento de atividades a distância em disciplinas de Estatística, por meio de ambientes virtuais de aprendizagem. O método inclui alguns recursos comumente encontrados em ambientes virtuais, como materiais de leitura, fórum, chat, correio eletrônico, FAQ, links, espaço compartilhado de trabalho e quadro de notas. O método foi aplicado a uma disciplina semipresencial do curso de graduação em Administração da FEA-RP/USP. Este trabalho se configura como uma pesquisa qualitativa e descritiva, com delineamento por estudo de caso único, cuja coleta de dados ocorreu por meio de observação participante (nos processos de tutoria), entrevistas do tipo focus group, relatórios de acessos dos alunos ao ambiente e um survey sobre o perfil da turma. O gerenciamento do curso ocorreu conjuntamente com o processo de tutoria da disciplina, o que permitiu acompanhar os acessos dos alunos a cada recurso do ambiente e o seu desempenho nas atividades propostas. Observou-se que alguns recursos não foram utilizados da forma planejada pelo método. A análise de regressão revelou que, durante a primeira fase da disciplina, as discussões no fórum e o desempenho nos trabalhos práticos influenciaram positivamente o desempenho do aluno na prova. Na segunda fase...

O aprendizado de Estatística vem se tornando cada vez mais importante para o sucesso de empresas que competem em cenário globalizado e tem destacado os profissionais que, de fato, possuem este conhecimento. Em especial na administração, a estatística auxilia nas tomadas de decisão e é usada como ferramenta em diversas áreas como Finanças, Marketing e Produção, contudo, é utilizada também em outros cursos. Sabe-se, porém, que existe certa dificuldade no aprendizado desta disciplina, seja porque a atitude perante a Estatística é negativa, ou mesmo porque existem diferenças de estilo de aprendizagem. Tanto aluno como professor têm a necessidade de conhecer melhor as variáveis que contribuem para o aprendizado efetivo. Com o objetivo de identificar como variáveis demográficas, estilos de aprendizagem dos alunos e a atitude perante estatística influenciam o desempenho dos alunos, este estudo analisou os as características, estilos, atitudes e desempenho dos alunos de Estatística aplicada da FEARP-USP. Foi utilizado o índice de estilos de aprendizagem de Felder e Soloman e a escala de atitudes de Schau. Os resultados desta análise reforçam os estudos que sugerem que as atitudes positivas influenciam no desempenho do aluno. Não foi conclusiva a associação entre gênero...

O foco de interesse deste estudo é a aprendizagem profissional de um grupo de professores e seus objetivos são: investigar como professores de Matemática da Escola Básica que pertencem a um grupo do tipo colaborativo problematizaram suas concepções sobre Educação Estatística nas práticas de ensinar e aprender Estatística; e compreender como o movimento do grupo possibilitou a sistematização de saberes profissionais dos professores. A questão de investigação ficou formulada da seguinte maneira: Como o movimento do grupo mobilizou práticas de ensinar e aprender Estatística e possibilitou a sistematização de saberes profissionais dos professores? A pesquisa é de natureza qualitativa, buscando uma abordagem histórico-dialética, em uma vertente interpretativa, procurando apreender o caráter dinâmico, contraditório e histórico dos fenômenos educativos. O trabalho de pesquisa, que teve a duração de um ano, se orienta segundo duas vertentes: o Desenvolvimento Profissional de Professores e a Educação Estatística. Foi fundamentado nos aportes teóricos dos estudos histórico-culturais de Bakhtin e na perspectiva de Investigação como Postura de Cochran Smith e Lytle, que trabalham as relações entre conhecimento e prática e o papel do conhecimento gerado pelos professores em suas práticas pedagógicas. O grupo do tipo colaborativo...

At a more technological and literate society statistics play a relevant role in order to allow people becoming critical and active citizens. Statistics is part of our daily life. Most media refer to statistical knowledge showing graphs, tables or means in order to sound scientifically supported and to be able to manipulate people’s opinion about what is going on in the world. Choosing the information one reads, analysing it, processing data and deciding different ways of presenting them are some of the competencies that we need to develop and mobilize. School practices play an essential role in the access pupils will have, or not, to these forms of literacy so deeply needed in a complex, changing and multicultural society. Nowadays most international and Portuguese policy documents refer that school should provide the means to develop each pupil’s competencies, namely the ones related to communication and to participant citizenship. Collaborative work is also suggested in many of them, namely related to statistical contents. Piaget and Vygotsky (Tryphon and Vonèche, 1996) underlined the role of communication in knowledge appropriation and in pupils’ performances. Social interaction, namely peer ones, played a main role in the process and could be seen as a facilitator when used within an innovative and coherent didactic contract (César...

peer-reviewed; While statistics is increasingly taught as part of the medical curriculum, it can be an unpopular subject and feedback from students indicates that some find it more difficult than other subjects. Understanding attitudes towards statistics on entry to graduate entry medical programmes is particularly important, given that many students may have been exposed to quantitative courses in their previous degree and hence bring preconceptions of their ability and interest to their medical education programme. The aim of this study therefore is to explore, for the first time, attitudes towards statistics of graduate entry medical students from a variety of backgrounds and focus on understanding the role of prior learning experiences.; PUBLISHED; peer-reviewed

It has recently been shown that supervised learning with the popular logistic loss is equivalent to optimizing the exponential loss over sufficient statistics about the class: Rademacher observations (rados). We first show that this unexpected equivalence can actually be generalized to other example / rado losses, with necessary and sufficient conditions for the equivalence, exemplified on four losses that bear popular names in various fields: exponential (boosting), mean-variance (finance), Linear Hinge (on-line learning), ReLU (deep learning), and unhinged (statistics). Second, we show that the generalization unveils a surprising new connection to regularized learning, and in particular a sufficient condition under which regularizing the loss over examples is equivalent to regularizing the rados (with Minkowski sums) in the equivalent rado loss. This brings simple and powerful rado-based learning algorithms for sparsity-controlling regularization, that we exemplify on a boosting algorithm for the regularized exponential rado-loss, which formally boosts over four types of regularization, including the popular ridge and lasso, and the recently coined slope --- we obtain the first proven boosting algorithm for this last regularization. Through our first contribution on the equivalence of rado and example-based losses...

An automated technique has recently been proposed to transfer learning in the hierarchical Bayesian optimization algorithm (hBOA) based on distance-based statistics. The technique enables practitioners to improve hBOA efficiency by collecting statistics from probabilistic models obtained in previous hBOA runs and using the obtained statistics to bias future hBOA runs on similar problems. The purpose of this paper is threefold: (1) test the technique on several classes of NP-complete problems, including MAXSAT, spin glasses and minimum vertex cover; (2) demonstrate that the technique is effective even when previous runs were done on problems of different size; (3) provide empirical evidence that combining transfer learning with other efficiency enhancement techniques can often yield nearly multiplicative speedups.; Comment: Accepted at Parallel Problem Solving from Nature (PPSN XII), 10 pages. arXiv admin note: substantial text overlap with arXiv:1201.2241

A learning environment, the tutor-web (http://tutor-web.net), has been developed and used for educational research. The system is accessible and free to use for anyone having access to the Web. It is based on open source software and the teaching material is licensed under the Creative Commons Attribution-ShareAlike License. The system has been used for computer-assisted education in statistics and mathematics. It offers a unique way to structure and link together teaching material and includes interactive quizzes with the primary purpose of increasing learning rather than mere evaluation. The system was used in a course on basic statistics in the University of Iceland, spring 2013. A randomized trial was conducted to investigate the difference in learning between students doing regular homework and students using the system. The difference between the groups was not found to be significant.; Comment: Presented at Edulearn 2013, Barcelona

The two-dimensional Gabor function is adapted to natural image statistics by learning the joint distribution of the Gabor function parameters. The joint distribution is then approximated to yield an analytical model of simple-cell receptive fields. Adapting a basis of Gabor functions is found to take an order of magnitude less computation than learning an equivalent non-parameterized basis. Derived learning rules are shown to be capable of adapting Gabor parameters to the statistics of images of man-made and natural environments. Learning is found to be most pronounced in three Gabor parameters that represent the size, aspect-ratio, and spatial frequency of the two-dimensional Gabor function. These three parameters are characterized by non-uniform marginal distributions with heavy tails -- most likely due to scale invariance in natural images -- and all three parameters are strongly correlated: resulting in a basis of multiscale Gabor functions with similar aspect-ratios, and size-dependent spatial frequencies. The Gabor orientation and phase parameters do not appear to gain anything from learning over natural images. Different tuning strategies are found by controlling learning through the Gabor parameter learning rates. Two opposing strategies include well-resolved orientation and well-resolved spatial frequency. On image reconstruction...

In many scientific settings data can be naturally partitioned into variable groupings called views. Common examples include environmental (1st view) and genetic information (2nd view) in ecological applications, chemical (1st view) and biological (2nd view) data in drug discovery. Multi-view data also occur in text analysis and proteomics applications where one view consists of a graph with observations as the vertices and a weighted measure of pairwise similarity between observations as the edges. Further, in several of these applications the observations can be partitioned into two sets, one where the response is observed (labeled) and the other where the response is not (unlabeled). The problem for simultaneously addressing viewed data and incorporating unlabeled observations in training is referred to as multi-view transductive learning. In this work we introduce and study a comprehensive generalized fixed point additive modeling framework for multi-view transductive learning, where any view is represented by a linear smoother. The problem of view selection is discussed using a generalized Akaike Information Criterion, which provides an approach for testing the contribution of each view. An efficient implementation is provided for fitting these models with both backfitting and local-scoring type algorithms adjusted to semi-supervised graph-based learning. The proposed technique is assessed on both synthetic and real data sets and is shown to be competitive to state-of-the-art co-training and graph-based techniques.; Comment: Published in at http://dx.doi.org/10.1214/08-AOAS202 the Annals of Applied Statistics (http://www.imstat.org/aoas/) by the Institute of Mathematical Statistics (http://www.imstat.org)

In this paper we study the problem of learning phylogenies and hidden Markov models. We call a Markov model nonsingular if all transition matrices have determinants bounded away from 0 (and 1). We highlight the role of the nonsingularity condition for the learning problem. Learning hidden Markov models without the nonsingularity condition is at least as hard as learning parity with noise, a well-known learning problem conjectured to be computationally hard. On the other hand, we give a polynomial-time algorithm for learning nonsingular phylogenies and hidden Markov models.; Comment: Published at http://dx.doi.org/10.1214/105051606000000024 in the Annals of Applied Probability (http://www.imstat.org/aap/) by the Institute of Mathematical Statistics (http://www.imstat.org)

EM algorithm is a convenient tool for maximum likelihood model fitting when the data are incomplete or when there are latent variables or hidden states. In this review article we explain that EM algorithm is a natural computational scheme for learning image templates of object categories where the learning is not fully supervised. We represent an image template by an active basis model, which is a linear composition of a selected set of localized, elongated and oriented wavelet elements that are allowed to slightly perturb their locations and orientations to account for the deformations of object shapes. The model can be easily learned when the objects in the training images are of the same pose, and appear at the same location and scale. This is often called supervised learning. In the situation where the objects may appear at different unknown locations, orientations and scales in the training images, we have to incorporate the unknown locations, orientations and scales as latent variables into the image generation process, and learn the template by EM-type algorithms. The E-step imputes the unknown locations, orientations and scales based on the currently learned template. This step can be considered self-supervision, which involves using the current template to recognize the objects in the training images. The M-step then relearns the template based on the imputed locations...

We consider the problem of approximating a given element $f$ from a Hilbert space $\mathcal{H}$ by means of greedy algorithms and the application of such procedures to the regression problem in statistical learning theory. We improve on the existing theory of convergence rates for both the orthogonal greedy algorithm and the relaxed greedy algorithm, as well as for the forward stepwise projection algorithm. For all these algorithms, we prove convergence results for a variety of function classes and not simply those that are related to the convex hull of the dictionary. We then show how these bounds for convergence rates lead to a new theory for the performance of greedy algorithms in learning. In particular, we build upon the results in [IEEE Trans. Inform. Theory 42 (1996) 2118--2132] to construct learning algorithms based on greedy approximations which are universally consistent and provide provable convergence rates for large classes of functions. The use of greedy algorithms in the context of learning is very appealing since it greatly reduces the computational burden when compared with standard model selection using general dictionaries.; Comment: Published in at http://dx.doi.org/10.1214/009053607000000631 the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)

The 2000 ASA Guidelines for Undergraduate Statistics majors aimed to provide guidance to programs with undergraduate degrees in statistics as to the content and skills that statistics majors should be learning. With new guidelines forthcoming, it is important to help programs develop an assessment cycle of evaluation. How do we know the students are learning what we want them to learn? How do we improve the program over time? The first step in this process is to translate the broader Guidelines into institution-specific measurable learning outcomes. This paper provides examples of how five programs did so for the 2000 Guidelines. We hope they serve as illustrative examples for programs moving forward with the new guidelines.

We consider the problem of learning causal information between random variables in directed acyclic graphs (DAGs) when allowing arbitrarily many latent and selection variables. The FCI (Fast Causal Inference) algorithm has been explicitly designed to infer conditional independence and causal information in such settings. However, FCI is computationally infeasible for large graphs. We therefore propose the new RFCI algorithm, which is much faster than FCI. In some situations the output of RFCI is slightly less informative, in particular with respect to conditional independence information. However, we prove that any causal information in the output of RFCI is correct in the asymptotic limit. We also define a class of graphs on which the outputs of FCI and RFCI are identical. We prove consistency of FCI and RFCI in sparse high-dimensional settings, and demonstrate in simulations that the estimation performances of the algorithms are very similar. All software is implemented in the R-package pcalg.; Comment: Published in at http://dx.doi.org/10.1214/11-AOS940 the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)

When dealing with datasets containing a billion instances or with simulations that require a supercomputer to execute, computational resources become part of the equation. We can improve the efficiency of learning and inference by exploiting their inherent statistical nature. We propose algorithms that exploit the redundancy of data relative to a model by subsampling data-cases for every update and reasoning about the uncertainty created in this process. In the context of learning we propose to test for the probability that a stochastically estimated gradient points more than 180 degrees in the wrong direction. In the context of MCMC sampling we use stochastic gradients to improve the efficiency of MCMC updates, and hypothesis tests based on adaptive mini-batches to decide whether to accept or reject a proposed parameter update. Finally, we argue that in the context of likelihood free MCMC one needs to store all the information revealed by all simulations, for instance in a Gaussian process. We conclude that Bayesian methods will remain to play a crucial role in the era of big data and big simulations, but only if we overcome a number of computational challenges.; Comment: Proceedings of the NIPS workshop on "Probabilistic Models for Big Data"

Many databases store data in relational format, with different types of entities and information about links between the entities. The field of statistical-relational learning (SRL) has developed a number of new statistical models for such data. In this paper we focus on learning class-level or first-order dependencies, which model the general database statistics over attributes of linked objects and links (e.g., the percentage of A grades given in computer science classes). Class-level statistical relationships are important in themselves, and they support applications like policy making, strategic planning, and query optimization. Most current SRL methods find class-level dependencies, but their main task is to support instance-level predictions about the attributes or links of specific entities. We focus only on class-level prediction, and describe algorithms for learning class-level models that are orders of magnitude faster for this task. Our algorithms learn Bayes nets with relational structure, leveraging the efficiency of single-table nonrelational Bayes net learners. An evaluation of our methods on three data sets shows that they are computationally feasible for realistic table sizes, and that the learned structures represent the statistical information in the databases well. After learning compiles the database statistics into a Bayes net...