How Machine Learning can be applied in a hospital environment and what are the most relevant issues encountered.
Classification algorithms tend to perform poorly when data is skewed towards one class, as is often the case when tackling medical diagnosis. A range of methods exist for addressing this problem, including under-sampling, over-sampling and ensemble methods.
Victor Vicente Palacios - Clinical Data Scientist @PhilipsSpain