Mineração de dados aplicada a fórum de discussão: uma proposta de visualização para gestores de educação a distância

• Autor(a) principal: Rodrigues, Fabiana Aparecida
• Data de Publicação: 2017
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações do Mackenzie
• Texto Completo: http://dspace.mackenzie.br/handle/10899/24452
• topic: ferramentas de comunicação; mineração de dados; redes neurais artificiais; educação a distância; fórum de discussão; visualização de dados; R studio; CNPQ::CIENCIAS EXATAS E DA TERRA

The mechanisms for managing evaluation and analysis of writing are strategies that may help to follow the tracks of information generated throughout e-learning courses. In this research, we detail the course of techniques application and presentation of results by means of reports. Such presentation was conditioned to the set of themes and to the evaluation criterion of Discussion Forum, which was restricted to the development of a set of analysis techniques about data generated in on-line courses interactions of a school institution, in order to enable the visualization and provide conditions to evaluate the quality of posts, as professors/tutors face di culties in following the discussions due to the quantity of interactions. For this reason, this study adopted as strategy a qualitative and quantitative approach that aimed to analyze both quantity of posts of 841 participants of 18 classes within the discussion forums universe where we may nd problems, and quality of students comments during the analysis, by checking if they were suitable and if they met the objective proposed. The objective of this research was extracting comments and having relevant information about the posts. To achieve this purpose, it was developed a set of techniques in an R Markdown le to present reports with results of the evaluation based on Data Mining (DM), speci cally Text Mining (TM) and exploratory analysis, whose investigations were summative and formative. In this way, we hope that the action developed in this research may support managers to analyze students contributions and follow the work of the professional in charge of class learning. The methodology showed and the indexes used may be applied in any type of text database.