A Preliminary Approach for Finding Linked Determinants between Diabetes and Caries using Genetic Algorithms

Abstract

Two of the public health conditions that continue having an increase in their incidence and prevalence worldwide are dental caries and diabetes mellitus. Dental caries is a multifactorial disease that affects the hard tissues of the tooth, being the most widespread disease in man; while diabetes mellitus is a multifactorial chronic disease that involves a group of disorders that alter the production and use of insulin, being of the main causes of disease in high, medium and low income countries. In this work is proposed the use of genetic algorithms in order to select the most significant features from a database of demographic data obtained from the NHANES 2013-2014, that best describe the difference between cases (diabetes and caries) and controls (healthy), according to their health status. After the feature selection, it is performed a validation stage, evaluating the multivariate model contained by these features. Within the results achieved shows a fitness > 0.980 and an AUC = 0.975, validating that the demographic features selected in this paper present a highly significant capacity to classify subjects that present caries and diabetes simultaneously from healthy subjects.