Random Forest Analysis of Combined Millimeter-wave and Near-infrared Sensing for Non-invasive Glucose Detection

Recently, great progress has been achieved in diabetes management through minimally invasive continuous glucose monitoring, which involves piercing the skin to identify changes in glucose levels in the interstitial fluid. Nevertheless, the development of accurate fully non-invasive glucose monitors remains a topic of great interest. This work examines the processing of data collected from a multi-sensor system, which combines millimeter-wave, near-infrared, and other auxiliary sensors' technology, coupled with a random forest machine learning model for prediction. We investigated possible correlations between millimeter-wave, near-infrared, skin temperature, pressure data, and blood glucose values, and deployed a machine learning model for prediction. The model is based on the random forest algorithm that was trained and tested on data collected from a study involving five healthy human subjects undergoing an intravenous glucose tolerance test. Our model achieves a root mean squared error of 21.06 mg/dl and a mean absolute relative difference of 7.31% for glucose prediction in the collected clinical subjects dataset. Additionally, 96.1% of the model's predictions fall within the clinically acceptable zones according to the Clarke error grid analysis. The preliminary results suggest that combining multiple sensor data with machine learning techniques is promising for the development of non-invasive glucose monitoring.