We developed a sensor-based health-monitoring system to observe the sweat rate using sensors. We used off-the-shelf and at-hand sensors to build our sweat-measuring system. One of the key reasons for using off-the-shelf sensors is to make the system available for DIY users and introduce it to the K 12 level students.
We presented an anomaly-based detection system using LightGBM, a fast and efficient machine learning algorithm. The advantages of LightGBM over other machine learning algorithms include low memory usage, higher accuracy, ability to handle large dataset, faster speed, and higher efficiency. CAN bus dataset can be huge the bus may use multiple baud rates up to 1 Mbit. Our proposed detection approach is based on LightGBM because algorithm is suitable for anomaly in CAN since the network is characterized by resource-constrained entities.