Prediction of Body Temperature from Smart Pillow by Machine Learning

Regression of Supervised Learning can predict continuous numerical values from labeled training data. A smart pillow consists of three temperature sensors, that can provide qualified data for regression algorithms to learn body temperature (BT). If BT can be extracted from a nonintrusive and unwired smart pillow accurately, it will be a big progress in the area of health care since BT is one of the most significant vital signs. Data are fed to adapted regression algorithms, such as K-Nearest Neighbor, Decision Tree, Random Forest, and Artificial Neural Network to predict BT, the results are compared to actual measured data, as well as the original Fuzzy Logic implementation. It proved that the application of Machine Learning algorithms can improve the accuracy of prediction for further analysis and development.