Machine Learning Based Non-Invasive Glucose Observation for Diabetes

In this paper, we discuss overcoming the traditional invasive technique to capture glucose levels and overcome this by using a non-invasive method to monitor glucose levels and other related parameters to give doctors a clear insight on diabetes. The clearer picture which we mention includes the implementation of Machine Learning algorithms for the early prediction and diagnosis of diabetes. The parameters include glucose levels, temperature, and heart rate and it may be very essential to reveal diverse clinical parameters. In modern healthcare systems, the use of IoT plays a vital role in the accessibility and monitoring of diverse patient data. The Internet of things serves as a catalyst for healthcare and performs an outstanding position in a huge variety of healthcare applications. In this venture, the microcontroller is used as a gateway to speak to the diverse sensors which include a temperature sensor and heartbeat sensor, glucose sensor. The microcontroller processes the sensor records and sends them to the cloud and subsequently presents the real-time data tracking of the parameters such as heart rate, temperature, and glucose levels for doctors. The records may be accessed each time with the aid. The records which are stored in the cloud are later used in Machine Learning algorithms to monitor the glucose levels and get a valuable prediction out of it for further monitoring purposes.