Result description
We developed a method based on machine learning that can create a model of user behaviour. Different types of behaviour can be modelled, but at the moment we model physical activity, sleep patterns and mental state, as obtained from sensors and questionnaires. This model can then predict likely future behaviour (e.g., increased or decreased physical activity) as well as detect anomalies in the current one (e.g., unusually low physical activity). Such predictions can in turn be used for monitoring the users’ progress, encouraging them or giving them personalised advice.
