We will develop an intelligent decision support system for optimal insulin dosing for Type I Diabetes mellitus management. The hypothesis used in this study is that case-based reasoning (CBR) methodology in the field of artificial intelligence (AI) can learn an individual's unique response to diet and exercise and provide a comprehensive explanation for its decision in each case; and that the self-organizing fuzzy neural network controller (SOFNNC) can use this information and its predictive control capability to determine the correct insulin dose input for maintaining the blood glucose level within an optimal range.In Phase I, as a proof-of-concept, a prototype system will be developed and evaluated, in response to changes in diet and exercise, for a group of well-selected and well- controlled patients. The intelligent insulin dosing system should enhance the ability of patients to balance diet, activity, and insulin. This will translate to better health and lower costs.National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK)
