Applying Predictive Data Mining to Discover Factors Associated to the Language Skill Performance from Elementary School Students

In this paper, predictive data mining techniques are applied to determine the academic performance from fifth grade students in Saber 5° tests Language skill at Colombian elementary schools 2017. We employed CRISP-DM methodology. Socioeconomic, academic, and institutional information was available ICFES databases. A minable dataset obtained using cleaning transformation techniques. decision tree built with Weka tool J48 algorithm. Some of predictors discovered patterns nature location school, whether or not failed a school year, age group, mother's educational attainment, rates ICTs household appliances. The findings research serve as quality for decision-making Ministry National Education (MEN) secretaries education, directors institutions define improvement plans that result education Colombia.