Aggregation model to measure the influence of machine learning and predictive models of employability relevant to the college career

Abstract

Higher Education Institutions (HEIs) train young professionals in technical and theoretical aspects, as well as competencies and skills in specific areas of knowledge, for a job market with very changing demands, which sometimes their graduates fail to meet expectations. This is why HEIs are constantly concerned about identifying the causes associated with the employability rate of their graduates. For this reason, HEIs must pay constant attention to the information they obtain in their graduate monitoring process, building a model to optimize the training process to calculate and estimate the HEI production scenarios based on the demand of the labor market. Graduate tracking systems must have Data Mining techniques that allow identifying graduate employability patterns for decision making in pursuit of a higher job insertion rate and a training of their professionals that responds to the labor demand, that is why the objective was raised to create and analyze three predictive learning models based on machine learning algorithms, which allow predicting the result of a target variable based on several predictor variables, the case studied is the Regional Autonomous University of the Andes, in Ecuador. The present research aims to develop an aggregation model to measure the influence of machine learning and predictive models of employability relevant to the university career. The predictive model used unsupervised learning algorithms, such as: linear regression, decision tree and SVM generating models with high precision and with the necessary adjustments demonstrating their applicability.