Clasificación de niveles de Medida de la Función Motora Gruesa mediante técnicas de Aprendizaje Automático

Abstract

The aim of the article is to classify the levels of gross motor function measurement (GMFCS) in minors using machine learning techniques. The study elements were 16 patients, boys, and girls between 2 and 9 years of age from a rehabilitation and physiotherapy institution suffering from cerebral palsy in gross motor function. The clinical analysis, the application of therapy and its measurement of gross motor function were collected, then the classification of nine machine learning algorithms was applied: k-Nearest Neighbor (k-NN), Gradient Boosted tree, Decision Stump, Random Tree, Rule Induction, Improved Neural Net, Generalized Linear Model, SVM, and Linear Discriminant Analysis, which were compared based on accuracy. The results obtained showed that the Linear Discriminant Model was the one that gave the best result with a 96.88 classification accuracy. Therefore, it is concluded that the use of machine learning techniques allows obtaining good accuracy in the classification of the measured level of gross motor function in boys and girls that can be used by specialists to carry out this task.