Description
Diabetes has been recognised as one of the leading cause of deaths around the world. Due to over-consumption of unhealthy foods and physical inactivity, the number of diabetic patients is raising drastically over the past few years. Proper treatment of diabetes needs measurements of Body Glucose Level (BGL) by laboratory tests or glucometers. These blood-based measurements are mostly invasive, costly, and often need experienced personnel to perform laboratory tests. As diabetes is associated with multiple disorders, there is indeed a great potential to identify different physical and physiological traits related to diabetes and utilise those for predictive modelling. The project aims to evaluate various physical, physiological, invasive, and non-invasive datasets – related to diabetes, propose a best performing dataset, and progress towards developing a cost-effective, easy-to-use Machine Learning (ML) model for its detection.
Requirements
Proficiency in programming language (Python preferred).
Basic understanding of statistics and ML model.
Background Literature
Cui, Ran, et al. "A Significance Assessment of Diabetes Diagnostic Biomarkers Using Machine Learning." Nurses and Midwives in the Digital Age. IOS Press, 2021. 36-38.
Susana, Ernia, et al. "Non-Invasive Classification of Blood Glucose Level for Early Detection Diabetes Based on Photoplethysmography Signal." Information 13.2 (2022): 59.
Daskalaki, Elena, et al. "The Potential of Current Noninvasive Wearable Technology for the Monitoring of Physiological Signals in the Management of Type 1 Diabetes: Literature Survey." Journal of medical Internet research 24.4 (2022): e28901.