Automated extraction of movement rationales for building agent-based models: Example of a Red Colobus monkey group

Abstract

The study of animal movement has gained impetus in recent years with improvements in telemetric technologies which enable high resolution tracking, providing researchers with a wealth of animal “big-data”. Coupling such movement data with information about the environments in which the animal moves provides a rich data source that can be exploited to understand an animal’s rationale for movement, which in turn can be used to extract “rules” that govern movement. The extraction of rules can be done using spatial, statistical and machine learning techniques. Once the rules replicating patterns and predictors of movement have been “discovered”, they can be subsequently used to build simulation models (ABMs) to mimic in-silico the behaviours of both individuals and groups of animals. We use field data collected by tracking Red Colobus (Procolobus rufomitratus) monkey groups from Kibale National Park, combined with land cover and terrain information, to show how this might be achieved.