Chapter 3 - Supervised and unsupervised learning

Abstract

Clinicians are constantly dealing with various information sources about our health, for example, ECG signals, brain activity, blood pressure, and oxygen saturation levels. However, due to the vast amount of data recorded and the limited resources, physicians tend to make treatment decisions based on isolated and short snapshots of a patient’s data. Machine learning offers automated analysis of heterogeneous measurement sources, helps detecting early warning signs of critical life-threatening events (e.g., heart attacks), improves clinical diagnostic accuracy, and helps clinicians make correct and timely medical interventions.