Implementation of LSTM for Prediction of Diabetes using CGM

Abstract

Deep learning has added conveniences for the diagnosis and prediction of various diseases making an influence in healthcare facilities. Diabetes mellitus is a dominant health issue faced by many around the globe. The number of people with this disease went up from one hundred eight million to six hundred in 1980, to four sixty million in 2019. Predicting trends of blood glucose prediction using deep learning methods make the management of the disease much easier. In this work, we are predicting future trends of the disease using training data. We have used the publically available dataset Ohio T1DM dataset in this work. In this paper, we have implemented LSTM to predict future trends. Root mean square error is used as the performance evaluation measure for this work.