Knowledge capture for the prediction and analysis of results of the quality test of higher education in Colombia

Abstract

In this paper, the prediction and results analysis of the quality test of higher education in Colombia - Saber Pro is performed. The knowledge extraction in databases methodology KDD was used, on which a database of the student's academic performance was built in areas associated with the contents of the Saber Pro test, and neural networks were used as a technique for data mining. The neural networks allowed the prediction of the results of the Saber Pro test with high exactness in both qualitative and quantitative ranges. A correlation between academic performance and Saber Pro results was also found. The findings suggest that the methodology used is an excellent guide for the discovery of hidden patterns in the data, and allows to establish strategies to improve the results of the Saber Pro tests that involve the student's academic performance.