Research on Diabetes Prediction Method Based on Machine Learning

Abstract

Diabetes mellitus is an inherited metabolism disorder described employing higher - blood sugar. This major medical type types one diabetes as well as type two diabetes. Presently, this generation for the younger generation suffering from type-one -diabetes has improved importantly. The type-one diabetes is prolonged whenever it occurs in adolescence also infancy, as well as has a long- incubation- period. These initial symptoms, in the beginning, are not clear, which might lead to failure or delaying treatment as well as the detection in time. Long-term higher- blood sugar may cause the especially eyes, kidneys, heart, blood vessels, and nerves, dysfunction of various tissues, as well as chronic damage. Thus, this initial prediction of diabetes is especially crucial. In the present study, we use managed machine-learning algorithms such as Naive Bayes classifier, Light-GBM also Support- Vector Machine (SVM) to instruct onto the actual data of potential diabetic patients aged sixteen to ninety as well as five-hundred twenty diabetic patients. In the comparative survey of the classification and recognition accuracy, the performance of the support vector machine is the best.