Advancing Statistical Methods for the Analysis of Complex Biologging Data Collected from Humans and Animals

Why Biologging Data?

Detecting underlying behaviours is an important goal of analyses of biologging data.

In medical sciences, biologging data are now being used to identify changes in heart rate variability and its association with mood changes in bipolar disorder.

Statistically, this goal can be translated into connecting observations to a finite set of latent states, with the additional challenge that biologging data typically exhibits high temporal dependence, such that observations close in time to one another are likely to be a product of the same underlying latent state.

Research Team’s Goals

• Using labelled data to improve state predictions
• Improving realism through the incorporation of feedback mechanisms
• Accounting for individuality in population inference
• Inference, identifiability, and model validation