Time series analysis of physiological data could potentially help healthcare professionals in decision making and even intervening to help to prevent dangerous clinical events. Commonly, time series analysis of physiological data is limited to a classification based approach. This is where only a single point or a class is predicted. The classification approach is useful but it does not provide much information. In contrast to this, it could be useful to explore and get further insight into what might be considered a dangerous clinical event. This is where accurate forecasting of physiological time series data could help to provide additional opportunities to explore the factors surrounding potentially dangerous clinical events. Furthermore accurate forecasting of physiological time series data is an open challenge. Another consideration is that few existing works that forecast physiological data have actually considered a comparison of the forecasting strategies taken and also the way that the data is included in the forecasting process.
Machine learning algorithms and methodologies specifically designed for forecasting of time series (as opposed to classification) have received little attention in this area, i.e. time series forecasting of physiological data. The aim of this project is therefore to investigate and explore the scope of time series forecasting of physiological data for the early intervention of dangerous clinical events through machine learning.