Intelligence Clinical Decision Support System for Diabetes Management

Abstract

Diabetes is the m ost common endocrine disease in all populations and all a ge groups. The diabetes patient should use correct therapy to live with this disease; there are several of important things to record about the patient and disease that help the doctors to make an optimal decision about the patient treatment. To improve the ability of the physicians, several tools have been proposed by the researchers for developing effective Clinical Decision Support System (CDSS), one of these tools is Artificial Neural Networks(ANN) that are computer paradigms that belong to the computational intelligence family. In this paper, a multilayer perceptron (MLP) feed-forward neural is used to develop a CDSS to determine the regimen type of diabetes management. The input layer of the system includes 25 input variables; the output layer contains one neuron that will produce a number that represents the treatment regimen. A Resilient back propagation (Rprop) algorithm is used to train the system. In particular, a 10-fold cross-validation scheme was used, an 88.5% classification accuracy from the experiments made on data taken from 228 patient medical records suffering from diabetes (type II).