Data mining applied to the risk of obesity in Cascavel/ PR

The aim of this study was to obtain patterns related to obesity and cardiovascular and comorbid risk factors in the city of Cascavel – PR, through the application of Data Mining (DM) techniques. Data about obesity were collected during the extension project of Unioeste’s Pharmacy graduation course, by means of a questionnaire with closed questions, which were analyzed in Weka program, and the results compared to the literature. According to the generated decisions tree, the obese individuals, with high abdominal circumference, have greater cardiovascular and comorbid risks. The best result of the model generated evaluation for the cardiovascular meta-risk attribute presented an accuracy of 94.8%, with 110 properly sorted instances, and for comorbid meta-risk attribute the accuracy was 84%. The DM generated useful informations for the decisions-making through the built data bank, assisting in the relevant parameter selection to obtain indicators of public health.