Educational DataMining: A Study of the Factors That Cause School Dropout in Higher Education Institutions in Brazil

Marília N. C. A. Lima, Roberta A. de A. Fagundes


Context:In Brazil, there is a high dropout rate in higher education institutions. Thus, it is clear that evasion is a frequent problem and that it is necessary to analyze the factors that cause it to enable solutions that can mitigate/ reduce this problem. Objetive: (1)perform a correlation analysis (Pearson and Spearman) of the educational factores of the School Census; (2)propose school dropout prediction models taking into account educational and economic factors using regression methods (linear, robust, ridge, lasso, clusterwise regression). Methodology: used the phases of the CRISP-DM methodology. Results: the factors related to not allowing financial assistance are related to as evasion, namely: food, permanence, didactic material, transportation. There are also factors related to the study period. The regression robust and linear regression show fewer errors. Conclusion: the correlations used present the selection of factors in a similar way, thus following a linear distribution. This study can help to create more investment in public policies, as it ratifies factors are related to this dropout problem.


Educational DataMining, Mineração de Dados Educacionais

Texto completo:



Direitos autorais 2020 RENOTE

RENOTE - Revista Novas Tecnologias na Educação      ISSN 1679-1916

A RENOTE, possui a nota Qualis B1 (Ensino) e B1 (Interdisciplinar) no Sistema de Avaliação da CAPES.

Centro Interdisciplinar de Novas Tecnologias na Educação (CINTED) - Universidade Federal do Rio Grande do Sul (UFRGS)

Avenida Paulo Gama, 110 - Anexo III – 3o andar
Porto Alegre, RS, Brasil – CEP: 90040-060

Telefone: (51) 3308-4100