DIAGNOSIS OF STUDENT CONFUSION THROUGH ARTIFICIAL INTELLIGENCE

Abstract

Student confusion is a problem that is experienced daily in school classrooms. Teachers transfer knowledge to students without knowing if they are receiving it correctly. Being aware of whether the student correctly grasps the knowledge would be of great help since different actions could be carried out to correct this problem. The proposal of this work focuses on carrying out the diagnosis of student confusion through Machine Learning algorithms. In particular, the following algorithms were applied: Logistic Regression (LR), [Formula: see text]-Nearest Neighbor (K-NN), Random Forest (RF) and Multi-Layer Perceptron (MLP). The metric was accuracy. The results obtained from the accuracy of each algorithm with the 5-Fold-Cross Validation validation method are 55.03% (LR), 52.98% (7-NN), 58.86% (RF) and 75.40% (MLP). An improvement in accuracy was achieved with respect to already published papers.