Verification of type 1 diabetes mellitus model

• Main Author: Sri Vathsavai Neelima
• Other Authors: Quek Hiok Chai
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence; DRNTU::Engineering::Computer science and engineering::Computing methodologies::Simulation and modeling
• Online Access: http://hdl.handle.net/10356/59104
• Institution: Nanyang Technological University

This study focuses on verifying Blood Glucose (BG) levels simulated using real patients input of meal and insulin details. This is important as diabetes pattern has been going up in the population over the last 10 years. It would be useful to have a personalized model for each diabetic patient to be able to predict supposed BG rise and falls so that diabetic patient can better manage their condition without constant supervision. This is the aim for the collaboration between KK Women’s and Children’s Hospital’s Endocrinology Centre and NTU’s School of Computer Engineering. This project is still in the early stages where T2DM (Type 2 Diabetes Mellitus) GlucoSim model has been tested against a real patients’ data. However, more data sets are needed to personalize this model to suit individual patients. For this, mass data collections over a period of 1 year with 3-4 months interval for the same patient should take place to better under the patient profile and general parameters unique to the patient. Upon data collection, data will be processed and entered into GlucoSim model to produce BG pattern. For this round of study, the focus has shifted to T1DM GlucoSim verification of 3 different patients. Based on the results of T1DM, the modelling is efficient in terms of predicting BG rise and fall with an estimate of 75% accuracy. However the parameters need to be further tuned for better results. For this, more studies and data is needed to derive the parameters.