Machine learning application in university management: Classification model Dropping out of engineering students in Peru

Abstract

We apply artificial intelligence to various cases of university management, with a proactive approach. In this study, we apply machine learning to classify whether a student will drop out or not, considering certain variables from the SITUATION DATA, based on the relevant attributes of the students. The results of this study would serve decision markers as a basis of inquiring about the high cost of not finishing the degree and adopting retention strategies. On the other hand, students can perceive the benefits of education, especially in engineering careers by having expectations, and value or assessment of it. The model developed is a neural network with 8 internal layers, in addition to the input and output layers, 225 training iterations have been considered, obtaining as a result an accuracy of 67.10%.