ABSTRACT
This chapter discusses how to use parameters in differential equations that model dynamic processes in time course data for classification. It focuses on convolutional network-based time course data analysis. The chapter introduces the function linear model with functional responses and functional predictors for quantitative genetic analysis of function-valued traits with next generation sequencing data. The wearable biosensors allow for the development of mobile health technologies that can continuously monitor patients, athletes, premature infants, children, psychiatric patients, people who need long-term care, elderly, and people in impassable regions far from health and medical services. Wearable biosensors allow continuous measurement of health-related physiology including electrocardiogram (ECG), seimocardiography, oxygen saturation levels, heart rate, skin temperature, blood pressure, and physical activities. The ECG is a time-varying signal which reflects the ionic current flow which causes the cardiac fibers to contract and subsequently relax. Deep learning has found large applications in genomics, wearable computing, disease prevention, diagnosis, and management.
