Article
5 Mathematical Background
Chronic kidney disease is a complex disease that impacts multiple organ systems throughout the body. Diagnosis and detection are often made through lab data collection and
analysis. Mathematical modeling can allow researchers to quantitatively represent multiple components and scales of a system and investigate the dynamic behavior of these
components and their interactions over time under various conditions. Thus, many researchers have been employing math modeling techniques to aid in CKD progress prediction such as analytical methods, numerical methods, and machine learning algorithms
9 Conclusions
We have managed to make significant headway in answering many of the fundamental questions that we set out to answer. Collections of patient lab data, biometric readings, symptoms, and baseline health factors were found to indicate deterioration of CKD. We correlated several metrics to eGFR and creatinine levels and identified several promising metrics to be used for remote monitoring when lab data isn’t available. A set of
health states were classified as being higher risk of CKD. Performance differences between various machine learning classifiers were provided, but we were unable to provide an analytical approach for which we would compare
our classifiers to.
As I an interested in AI and machine learning, research on feline kidney disease led me to this cambridge research paper on using machine learning to predict CKD progression. Seeing how these new technologies can be worked with veterinary medicine surprised me and I would like to research more on this link between machine learning and CKDs.
