Utilização de aprendizagem de máquina para a identificação de dependência em aparelhos celulares com foco em casos que possam causar reprovação e evasão

Abstract

This work introduces the problems caused by the excessive use of cell phones in modern society and shows that they can cause school problems to students who use it too much. For this, we integrated technology with the use of an artificial intelligence area, machine learning, with data from a specific database. For the proposed task, the Naive Bayes, AdaBoost, SVM, Bagging and Random Forest classifiers were used. At the end of the tests, the SVM classifier presented the best performance in the general context. With rates ranging from 91.22 to 80 percent in the three scenarios tested.