Predição do risco de evasão dos cursos técnicos a distância do IFPI : uma aplicação da metodologia CRISP-DM em dados educacionais

• Autor(a) principal: PÁDUA, Antonio Francisco Lima de Oliveira
• Data de Publicação: 2017
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UFRPE
• Texto Completo: http://www.tede2.ufrpe.br:8080/tede2/handle/tede2/7950
• topic: Evasão escolar; Educação a distância; Ensino técnico; CIENCIAS HUMANAS::EDUCACAO

This study aims to understand the contribution of knowledge discovery in educational databases through the application of the Cross Industry Standard Process for Data Mining (CRISP-DM) methodology to the problem of school dropout from distance learning courses. Thus, the main goal was to adapt a data mining process to estimate the evasion risk of students from Secondary Technical Level Courses, in the Distance Learning modality, on the Federal Institute of Piauí (IFPI), under the aforementioned methodology. To do so, it was necessary to understand the knowledge discovery over high volume data and the CRISP-DM methodology as a paradigm of educational data mining; To apply the CRISP-DM methodology to identify the common characteristics of students who drop out of the technical distance courses of the IFPI, in order to allow the management coordination a chance to proceed with preventive intervention; And to evaluate the risk of evasion of students from technical courses in the Distance Learning modality on the IFPI, based on academic data of incoming students from the year 2013. The research has quantitative, documental, descriptive and applied approaches. We used the first five phases of the CRISP-DM methodology in data collected from the socioeconomic questionnaires, and the final situation of the students regarding the classification of evaded or not evaded after the first semester of the Distance Learning Technical Course in Work Safety of the Federal Institute of Education, Science and Technology of Piauí, Campus Teresina Zona Sul. Data collection was done through Google form and the data integration was done using the Excel 2010 software. In order to identify the predictor variables that explain the target variable, school dropout, we used logistic regression through the Ordinary Least Squares (OLS) function of Gretl 2016c software. In light of this study, it is possible to say that the use of data mining techniques following the CRISP-DM methodology based on distance learning courses can also contribute to courses in any modalities regarding the search for strategies against school dropout.