MACHINE LEARNING APPLIED TO ICFES TESTS TO IDENTIFY MEASUREMENT PATTERNS IN STUDENTS

In this paper, we will analyze the different techniques of Machine Learning in higher education through the information stored in the Colombian Institute for the Evaluation of Education in Colombia (Icfes) to measure the most relevant characteristics for a student to get both positive and negative results. For this, we will use algorithms based on the Regression Model, known as supervised algorithms: Naïve Bayes and K Nearest Neighbors, whose functions are to determine through the distance between two points, which can be related and the selection of variables, taking into account the class variable and its dependencies on the data. The standardized tests were created in 2009, and since then, they have collected on average more than one million data, detailed enough to know general aspects such as family, origin, school, assets, among others, and specific ones (qualifications). Combining the information from recent years will make it possible for us to analyze the progress on education and what extent of the most deficient variables should be improved to help students become holistic human beings.