No description available.

¡@¡@e-Learning is becoming more and more important for many educational institutions, and many educators believe that there is a good potential for providing adaptive learning in e-learning environment. In order to support the design of adaptive learning materials, teachers need to refer to the ontology of the subject domain to be taught. Moreover, ontology can show the whole picture and the core knowledge of a subject domain. Literature reviews also pointed out that graphical representation of ontology can reduce the problems of information overloading and learning disorientation for learners. However, ontology constructions all rely on domain experts in the past; it is a time consuming and high cost task. It would be more challenge for those emerging new domains like e-Learning. ¡@¡@e-Learning is a new and fast developing domain, how to automatic constructing its ontology is a very important topic. In this research, we use some relevant e-Learning journals and conferences papers as input data sources, and apply data mining techniques to automatically construct the concept maps for e-learning. We also analyzed the evolution in e-Learning domain according to the concept maps constructed at different time periods. ¡@¡@The contribution of this research is automatic constructing the concept maps of e-Learning domain using text-mining techniques. It can provide a comprehensive and useful reference for researchers to do research, for teachers to do adaptive course design and for learners to understand the related knowledge in e-Learning.

Remotely-sensed hyperspectral imagery has become one the most advanced tools for analyzing the processes that shape the Earth and other planets. Effective, rapid analysis of high-volume, high-dimensional hyperspectral image data sets demands efficient, automated techniques to identify signatures of known materials in such imagery. In this thesis, we develop a framework for automatic material identification in hyperspectral imagery using adaptive similarity measures. We frame the material identification problem as a multiclass similarity-based classification problem, where our goal is to predict material labels for unlabeled target spectra based upon their similarities to source spectra with known material labels. As differences in capture conditions affect the spectral representations of materials, we divide the material identification problem into intra-domain (i.e., source and target spectra captured under identical conditions) and inter-domain (i.e., source and target spectra captured under different conditions) settings. The first component of this thesis develops adaptive similarity measures for intra-domain settings that measure the relevance of spectral features to the given classification task using small amounts of labeled data. We propose a technique based on multiclass Linear Discriminant Analysis (LDA) that combines several distinct similarity measures into a single hybrid measure capturing the strengths of each of the individual measures. We also provide a comparative survey of techniques for low-rank Mahalanobis metric learning, and demonstrate that regularized LDA yields competitive results to the state-of-the-art, at substantially lower computational cost. The second component of this thesis shifts the focus to inter-domain settings, and proposes a multiclass domain adaptation framework that reconciles systematic differences between spectra captured under similar, but not identical, conditions. Our framework computes a similarity-based mapping that captures structured, relative relationships between classes shared between source and target domains, allowing us apply a classifier trained using labeled source spectra to classify target spectra. We demonstrate improved domain adaptation accuracy in comparison to recently-proposed multitask learning and manifold alignment techniques in several case studies involving state-of-the-art synthetic and real-world hyperspectral imagery.

The security of a steganography system is defined by our ability to detect it. It is of no surprise then that steganography and steganalysis both depend heavily on the accuracy and robustness of our detectors. This is especially true when real-world data is considered, due to its heterogeneity. The difficulty of such data manifests itself in a penalty that has periodically been reported to affect the performance of detectors built on binary classifiers; this is known as cover source mismatch. It remains unclear how the performance drop that is associated with cover source mismatch is mitigated or even measured. In this thesis we aim to show a robust methodology to empirically measure its effects on the detection accuracy of steganalysis classifiers. Some basic machine-learning based methods, which take their origin in domain adaptation, are proposed to counter it. Specifically, we test two hypotheses through an empirical investigation. First, that linear classifiers are more robust than non-linear classifiers to cover source mismatch in real-world data and, second, that linear classifiers are so robust that given sufficiently large mismatched training data they can equal the performance of any classifier trained on small matched data. With the help of theory we draw several nontrivial conclusions based on our results. The penalty from cover source mismatch may, in fact, be a combination of two types of error; estimation error and adaptation error. We show that relatedness between training and test data, as well as the choice of classifier, both have an impact on adaptation error, which, as we argue, ultimately defines a detector's robustness. This provides a novel framework for reasoning about what is required to improve the robustness of steganalysis detectors. Whilst our empirical results may be viewed as the first step towards this goal, we show that our approach provides clear advantages over earlier methods. To our knowledge this is the first study of this scale and structure.

No description available.

Submitted by Mariana Dornelles Vargas (marianadv) on 2015-05-25T12:29:15Z No. of bitstreams: 1 AutoEduMat.pdf: 1060362 bytes, checksum: 25b8156de4b9c2c2c5b9dc0f69aea011 (MD5) / Made available in DSpace on 2015-05-25T12:29:15Z (GMT). No. of bitstreams: 1 AutoEduMat.pdf: 1060362 bytes, checksum: 25b8156de4b9c2c2c5b9dc0f69aea011 (MD5) Previous issue date: 2010 / Nenhuma / Esta dissertação apresenta uma pesquisa relacionada as ferramentas que dão suporte a utilização de objetos de aprendizagem em plataformas digitais. Mais especificamente, a pesquisa se direciona para as ferramentas de apoio a autoria destes objetos, em particular dos seus metadados. Inicialmente é apresentada a contextualização do problema de pesquisa, sua fundamentação teórica e os trabalhos relacionados ao tema. Em seguida são apresentadas as principais características do sistema proposto, o AutoEduMat - Ferramenta de Apoio a Autoria de Metadados de Objetos de Aprendizagem para o Domínio de Ensino de Matemática. A ferramenta AutoEduMat dá apoio a autoria de objetos de aprendizagem, oferecendo assistência ao projetista (designer) de objetos na criação e edição de metadados destes objetos. A principal inovação do trabalho é a combinação das tecnologias de Engenharia de Software de Agentes e de Engenharia de Ontologias para construir um sistema multiagente que oferece suporte inteligente para a geração dos metadados dos objetos de aprendizagem, sendo capaz de interagir com o usuário com termos de seu próprio contexto profissional e educacional. No trabalho é proposta a ontologia Onto-EduMat que incorpora os conhecimentos sobre o domínio de ensino de matemática, incluindo aspectos pedagógicos, necessários para o auxílio a geração dos metadados. Tanto a ferramenta quanto seu modelo ontológico são validados através de experimentos descritos no final do trabalho. / This dissertation presents a research related to the tools that support the utilization of learning objects in digital platforms. More precisely, the research is directed to the tools that support the authoring process of these objects, in particular of their metadata. Initially are presented the characterization of the problem, its theoretical foundations and related works. Then are presented the main characteristics of the proposed system, the AutoEduMat - Metadata Authoring Tool for Mathematics Learning Objects. The AutoEduMat system will provide assistance to the object designer in the metadata creation and edition of these objects. The main innovation of this work is the combination of Agent Oriented Software Engineering and Ontology Engineering technologies to built a multiagent system able to offer intelligent support for metadata creation, interacting with users using terms related to their professional and educational context. This work proposes the Onto-EduMat ontology, which incorporates the mathematical and pedagogical knowledge necessary to generate the metadata. The authoring tool and its ontological model are validated through experiments described in the end of the work.