Sistema de apoyo para la acreditación de la calidad de programas académicos de la Universidad de Caldas, aplicando técnicas en minería de datos

One of the main difficulties faced by the current educational system is desertion, its accumulated value reaches levels of 45% at the national level and 33% at the University of Caldas; one of its main causes is academic dropout, which is why the definition of an indicator that manages to measure academic performance in its different dimensions (excellence, efficiency and effectiveness) is essential. Once the indicator has been defined, it is essential to determine the factors that affect it in order to take actions that have the greatest possible impact; These factors include elements of identity, socioeconomic, vocational, previous studies, the family environment among others and their relationships, so a multivariate analysis should be the type of model that characterizes them, logistic regression and its wide use in determining of risk or protective factors was the selected analysis, both for this characteristic and for its handling of numerical and categorical variables. This process of knowledge extraction from KDD data, which has been on the rise in recent years in educational environments, was developed using the CRISP-DM methodology and was implemented on the free-use application RapidMiner and includes from the extraction of information from the SIA Academic Information System database, its transformation, validation, the calculation of the indices and the IAR academic performance indicator, the Logistic Regression analysis by program and level of progress in program credits, up to the generation of descriptive reports such as the model. The final results show the risk and protective factors in academic performance for students in each face-to-face program at different times during their time at university.