This work presents an analysis and a classification of publications in the field of Educational Data Mining published by Brazilian researchers in events and periodicals in Latin America over the past four years. The objective of the present work is to point and present what is being researched in academic performance prediction area. The results of the classification of publications are presented and we show interesting information about the area of research, identifying some possibilities less (or not yet) explored.Resumo. O presente trabalho apresenta uma análise e classificação de publicações na área de Mineração de Dados Educacionais publicados por pesquisadores brasileiros em eventos e periódicos na América Latina nos últimos quatro anos. O objetivo deste trabalho é apontar e apresentar o que está sendo pesquisado na área de predição de desempenho acadêmico. Os resultados da classificação das publicações são apresentados e mostram informações interessantes a respeito da área de pesquisa, identificando possibilidades pouco (ou ainda não) exploradas.
The academic performance prediction can be very useful for Educational Institutions in order to help them to take pedagogical decisions that can help students. In this work, we present experiments using Moodle data, Time Series and the Feature Selection Wrapper approach, since, to best of our knowledge, this method to reduce the number of features have not been used in this "kind" of data. Results showed an improvement in the performance of classifiers, some obtaining the mark of 84.7% in accuracy results.
Resumo.A previsão de desempenho acadêmico tem grande utilidade para Instituições de Ensino no sentido de auxiliá-las a tomar, de forma antecipada, decisões pedagógicas que possam auxiliar os estudantes. Neste trabalho foram realizados experimentos em uma base de dados do Ambiente Virtual de Aprendizagem Moodle, utilizando o conceito de Séries Temporais e a técnica Cápsula de Seleção de Atributos que, dentre os trabalhos pesquisados, não havia sido ainda empregada. Resultados experimentais indicam uma melhora no desempenho dos classificadores com o uso de Seleção de Atributos, alguns alcançando a marca de 84,7% de acurácia.
Authoring tools for hypermedia languages usually provide visual abstractions, which hide the source code from the author aiming to simplify and accelerate the development process. Among other drawbacks, these abstractions modify or even break the communication process between the author and the language designer, since these languages were designed to be readable and understandable by its target audience. This paper presents a textual approach to hypermedia authoring that does not have these inconveniences, but rather uses typographical accessories, such as program visualization, hypertextual navigation, and semiautomatic error correction. The proposed approach exploits concepts known to the author and does not imply in extra cognitive overload. A use case is presented, namely the NCL Eclipse authoring environment, for Nested Context Language, the Brazilian Digital TV and ITU-T standard.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.