Modelling the Movement of a South African Cheetah Using a Hidden Markov Model and Circular-Linear Regression

Cheetah numbers are decreasing drastically and analysing these types of datasets can aid in the efforts to conserve this species. Studying GPS data gives insight into the habitats and migratory patterns of animals. In this paper, firstly, a generalised hidden Markov model is explored for modelling the movement of Rain, a female cheetah in Pilanesberg National Park, South Africa to decode the most probable hidden states (which include encamped and exploratory). Secondly, circular-linear regression is applied to predict the direction of her movement in the presence of linear predictor variables. The study of animal movement can help answer many ecological questions, such as in this case how human encampment can influence the behaviour of large predators.