Resumo

FONSECA, STELLA OGGIONI DA et al. DATA MINING GUIDED BY THE EDUCATIONAL DOMAIN: A PROOF OF CONCEPT. Est. Aval. Educ. [online]. 2019, vol.30, n.74, pp.420-453.  Epub 22-Out-2019. ISSN 1984-932X.  https://doi.org/10.18222/eae.v30i74.5696.

This paper aims to propose a methodology for identifying patterns related to learning mathematics and the characteristics of the school environment. It was applied to the data of the Prova Brasil 2013, focusing on students in their 9 th year of education, in the state of Rio de Janeiro. The approach, supported by education experts, consists of proposing a dimensionality reduction process integrated to data mining. Its evaluation was conducted using technical measures together with measures of interest to the educational domain. This approach made it possible to identify actions, as well as to analyze their viability, to solve educational issues. The methodology, guided by the area of application and the evaluation of which is not restricted to the use of technical metrics of data mining, can serve as a reference - a proof of concept - for other studies of educational actions and policies.

Palavras-chave : PROVA BRASIL; DATA MINING; EDUCATIONAL POLICIES; SCHOOL PERFORMANCE..

        · resumo em Português | Espanhol     · texto em Português     · Português ( pdf )