Predictive Analytics for Chronic Diabetes Care

摘要

In this paper, we investigate diabetic patient classification using Multilayer Perceptron (MLP) and Bayesian Networks (BN). The models were evaluated on the basis of accuracy, Root Mean Squared Error (RMSE), and Area under Receiver Operating Curve (AUC). The dataset used for evaluation was CDC-NHANES 2011-2012, which is a survey of approximately 5,000 individuals' health and nutritional status as collected by the Centers for Disease Control on a yearly basis. In all classification instances, MLP networks were shown to have a larger AUC, higher accuracy, and lower RMSE.