Análise de desempenho discente em ambiente virtual de aprendizagem: mineração de dados educacionais através do processo CRISP-DM

• Autor(a) principal: Fonseca, Enir da Silva
• Data de Publicação: 2020
• Tipo de documento: Tese
• Idioma: por
• Título da fonte: Repositório do Centro Universitário Braz Cubas
• Texto Completo: https://repositorio.cruzeirodosul.edu.br/handle/123456789/3237
• topic: Análise de dados; Descoberta de conhecimento; Processo CRISPDM; Mineração de Dados Educacionais; Tomada de decisão.; CNPQ::CIENCIAS HUMANAS::EDUCACAO::ENSINO-APRENDIZAGEM

The continuous technological evolution, has provided an increase in structured and unstructured data, generated from varied sources and different forms. And for the resolution of a certain problem, the correct interpretation is necessary for decision making, challenges that accompany society. To meet the main objective of this research, which is to analyze development and performance and student guided by data interpretation and analysis using the CRISP-DM process, data mining resources were used with the CRISP-DM process and an analysis of the data using RStudio software, which is a development environment integrated with R, a programming language for graphs and statistical calculations. To achieve this, the data was investigated in a virtual learning environment configured on the Blackboard, collecting information about the user's activities in the forums, all the user's activity in the content areas, their respective notes in the VLE and the result after the face-to-face evaluations. The consultation took place in 9 subjects, divided between 19 classes offered in 2018, with a record of activities of 7230 students. After the analysis, it was identified that 77.23% of the students obtained a final average equal to more than 7 points, and that the best grades are associated with the group with access to the virtual learning environment, which was 160 times higher until the time of the evaluations. in person. The students who obtain in the VLE, a grade equal to or higher than 3 points, tend to reach the maximum grade in the final evaluation. With a two-dimensional analysis, a moderate and positive linear relationship between the variables was verified, the concentration of accesses in the VLE contributes to obtaining the student's final average. Thus, the study proved to be satisfactory for data analysis and decision making, enabling feedback and other analyzes during the operation, with the feedback of new data and specific corrections before the decision-making process.