The Prediction Of Blood Glucose Level By Using The ECG Sensor of Smartwatches

The objective of this research paper is to try out different machine learning models on the patients from the D1NAMO data-set in order to predict blood glucose level based on the heart rate values along with extra features coming from the smart watch, The glucose readings will be measured by the CGM along the 2 weeks duration of the calibration between smartwatch and CGM device. Our proposed System will be applied per patient, the patient will use the CGM device along with the smart watch and calibrate the 2 readings together, after the CGM expires the data will be taken and processed together, any noise that interfered with the ECG signal will be removed in order to have a clean noise that can be combined with the Glucose readings to be in the exact timestamp, as an output the user will have the ability to depend only on the smart watch in knowing the blood sugar level for a duration of time.