Using the Belbin method and models for predicting the academic performance of engineering students

This paper describes the process of generating a predictive model of students' academic performance in different engineering subjects at Universidad Catolica del Norte (UCN). It aims to analyze the importance of variables influencing the final average grade of the UCN students in projects related to different subjects, focusing on the dimensions resulting from the Belbin test. The main objective of this work is to provide evidence of the real impact of the Belbin test outcomes on the final performance of a team of students, using as a metric of variable importance the one provided by a Random Forest model, supplied by the scikit-learn library. As a result, the final classifier presented an accuracy of 80%, and one of the most influential variables according to this model was Covered Roles 2, which represents the number of roles covered in each group. Future research lines are proposed to validate these outcomes, mostly concerned with the acquisition of more data across several future semesters.