Detección de patrones de desempeño académico en la competencia de matemáticas en las pruebas Saber 5°.

The Saber 5° tests seek to contribute to improving the quality of education in Colombia. Their application is periodic, they evaluate students' basic skills, and they analyze the factors that affect their achievements. This paper presents one of the results of a research work whose objective was to apply data mining techniques to detect patterns of academic performance regarding the mathematics component of the Saber 5 tests in the year 2017. This test was taken by the fifth-grade students of Colombian Primary Education institutions. To meet this objective, the CRISP-DM methodology was used. Socioeconomic, academic, and institutional information was obtained from the ICFES databases. This information was preprocessed using data cleaning and transformation techniques. The decision tree classification model was selected, as it allows easily interpreting patterns. Among the most important factors in the patterns associated with good or poor academic performance in mathematics are the nature and location of the school and whether or not the student had failed a grade. The knowledge generated in this research constitutes quality information for decision-making by the Ministry of National Education, education secretariats, and the executives of Primary Education institutions with regard to the definition of improvement plans that result in the quality of Primary Education in Colombia. [ABSTRACT FROM AUTHOR]