Estimation of Gestational Diabetes Mellitus using Azure AI Services

Abstract

Diabetes is a condition that slowly affects the natural defense mechanism and vital organs; it's a silent and slow killer. It is a condition when body fails to utilize the energy provided by food due to excessive blood sugar level. Diabetes is among the potentials reasons for kidney diseases, heart-stroke, blindness and lower limb amputation. WHO estimates that more than 500 million people are infected with diabetes worldwide, the disease is responsible for over more than 2.5 million causalities across the globe, the condition is more prevalent in medium and low income nations. Paper presents a procedural approach for Estimation of Gestational Diabetes using Microsoft Azure AI services. Microsoft Azure AI is a Machine Learning Studio, with all major functions and algorithm at the view of drag and drop. The classification model building the work, can predict the occurrence of Gestational diabetes based on factors at early stages of pregnancy. Pima Indian dataset from UCI has been utilized to build and cross validated the predictive model. The algorithm was examined on a dataset of 768 samples and it was able to achieve an accuracy of 77.8%.