First Experiences with the Identification of People at Risk for Diabetes in Argentina Using Machine Learning Techniques

Detecting Type 2 Diabetes (T2D) and Prediabetes (PD) is a real challenge formedicine due to the absence of pathogenic symptoms and the lack of knownassociated risk factors. Even though some proposals for machine learning modelsenable the identification of people at risk, the nature of the condition makesit so that a model suitable for one population may not necessarily be suitablefor another. In this article, the development and assessment of predictivemodels to identify people at risk for T2D and PD specifically in Argentina arediscussed. First, the database was thoroughly preprocessed and three specificdatasets were generated considering a compromise between the number of recordsand the amount of available variables. After applying 5 differentclassification models, the results obtained show that a very good performancewas observed for two datasets with some of these models. In particular, RF, DT,and ANN demonstrated great classification power, with good values for themetrics under consideration. Given the lack of this type of tool in Argentina,this work represents the first step towards the development of moresophisticated models.