| Summary: | Submitted by Setor de Tratamento da Informa??o - BC/PUCRS (tede2@pucrs.br) on 2016-12-30T11:17:36Z
No. of bitstreams: 1
DIS_AUGUSTO_WEIAND_COMPLETO.pdf: 3232365 bytes, checksum: 2e18edf9e1a20e9351da458aacc8144d (MD5) === Made available in DSpace on 2016-12-30T11:17:37Z (GMT). No. of bitstreams: 1
DIS_AUGUSTO_WEIAND_COMPLETO.pdf: 3232365 bytes, checksum: 2e18edf9e1a20e9351da458aacc8144d (MD5)
Previous issue date: 2016-03-01 === With the technology advancement, distance education has been very discussed in recent years, especially with the emergence of several kinds of Virtual Learning Environments (VLE?s). These environments used in distance education courses, usually generate a lot of data due to the high number of students and the various tasks which involve their interactions. Thus, arises the need to search efficient and intelligent ways to find relevant information. Data mining techniques help in the discovery of implicit knowledge that can support decision making. However, eventually appear difficulties in understanding the obtained results of the mining due to the analyzed volume. In these cases, the use of visualization and interaction techniques assists in this task. The main goal of this work is to present the development of a visual analysis approach that uses data mining algorithms and visualization techniques to help monitoring students of distance learning courses in the institutions that use virtual learning environments. These students are classified considering their performance, providing ways to investigate and predict possible approvals, disapprovals and evasions. The visualizations aim to improve the understanding of the generated data by the mining algorithms, providing different ways of interaction. It is possible to analyze both the general behavior of students in a selected course, as their individual behaviors. Performance comparisons of a student between different courses, and from interactions performed in a set of courses are also allowed. Initial tests demonstrated that it was possible to make predictions in a satisfactory way, as well as enable visualizations and interactions to the users for interpreting the information resulting from mining algorithms. === Com o avan?o da tecnologia, a educa??o a dist?ncia tem sido muito discutida nos ?ltimos anos, especialmente com o surgimento de diversos tipos de Ambientes Virtuais de Aprendizagem (AVA?s). Estes ambientes, quando utilizados em cursos de educa??o a dist?ncia, normalmente geram uma grande quantidade de dados devido ao elevado n?mero de alunos e as diversas tarefas que envolvem as suas intera??es. T?cnicas de minera??o de dados auxiliam na descoberta de conhecimentos impl?citos que possibilitem dar suporte ? tomada de decis?o. Por?m, eventualmente surgem dificuldades no entendimento dos resultados obtidos pela minera??o, devido ao volume de dados analisado. Neste caso, o uso de t?cnicas de visualiza??o e intera??o auxiliam nesta tarefa. Este trabalho tem como objetivo apresentar o desenvolvimento de uma abordagem de an?lise visual, que utiliza algoritmos de minera??o de dados e t?cnicas de visualiza??o para auxiliar no acompanhamento de alunos de cursos a dist?ncia nas institui??es que utilizam ambientes virtuais de aprendizado. Estes alunos s?o classificados considerando o seu desempenho, possibilitando a investiga??o e predi??o de poss?veis aprova??es, reprova??es ou evas?es. ? poss?vel analisar tanto o comportamento geral dos alunos de uma disciplina selecionada, como seus comportamentos individuais. Compara??es de desempenho de um aluno entre diferentes disciplinas, e das intera??es realizadas em um conjunto de disciplinas tamb?m s?o permitidas. Testes iniciais demonstraram que foi poss?vel efetuar previs?es de maneira satisfat?ria, assim como, possibilitar aos usu?rios visualiza??es e intera??es para interpretar estas informa??es advindas dos algoritmos de minera??o.
|
|---|
