Machine Learning models based in Supervised Learning for the Detection of Diabetes Mellitus in the City of Guayaquil

Authors

Patiño-Pérez, Darwin

Iñiguez Muñoz, Felicisimo

Nivela Cornejo, María

Ruíz Ramírez, Alicia

Otero Agreda, Omar

Game Mendoza, Karla

Vinueza Burgos, Gladis

Abstract

Due to the problems presented by public health centers in the city of Guayaquil in Ecuador, a group of specialists greeted the creation of a comprehensive diabetes diagnosis center, which, with an entrepreneurial initiative, is bringing together private health centers that are interested in providing a technologically assisted service. Using artificial intelligence tools, the comprehensive center detects whether a person has diabetes mellitus or not; Therefore, through this mechanism and applying the established triage, the person detected by the AI model is referred to the respective private health center that is part of the comprehensive center (entrepreneurship). Supervised learning was applied in the implementation of some conventional machine learning techniques, which were compared with an artificial neural network model to determine the model that can obtain the highest accuracy in learning diabetes diagnosis. As a result, it was obtained that the artificial neural network ANN reached 99.33% accuracy compared to the support vector machines SVM, which reached 98.6% accuracy in the classification of patients with diabetes mellitus, so that in view of the fact that the RNA model , is the one who has learned to detect diabetes with greater accuracy and is recommended to be used in the comprehensive diagnostic center, for rapid referral to a private health center that can apply rapid and effective treatment .