Skip to main content

behavioural change point analysis (BCPA)

A novel method for identifying behavioural changes in animal movement data, could be useful for migration/resident catagorization. some starter r-code included

"Here we develop a novel, robust and efficient method for identifying behavioural changes in behaviourally heterogeneous and temporally gappy movement data without any prior assumptions. Step lengths and turning angles are transformed into orthogonal persistence and turning components of velocity and characterized as continuous autocorrelated time series described locally by three parameters: a mean, a variance and a continuous autocorrelation."

"By sweeping an analysis window over an entire movement path, an aggregated behavioural summary of movement is obtained. The complete suite of steps is termed a behavioural change point analysis (BCPA)."

"The fundamental contribution of the BCPA is the ability to detect and characterize significant behavioural shifts without a priori assumptions. The framework permits not only the identification of discrete shifts, but also the detection of gradual changes in the parameter values. While similar movement analyses have presupposed a few distinct behaviours and classified behaviour into categories (Morales et al. 2004; Forester et al. 2007; Jonsen et al. 2007; Bailey et al. 2008), the BCPA reveals striking complexity in behavioural modes, both within a single movement track and in comparisons between tracks of a single individual"

Gurarie et al 2009 A novel method for identifying behavioural changes in animal movement data