Behaviour Classification on Giraffes (Giraffa camelopardalis) Using Machine Learning Algorithms on Triaxial Acceleration Data of Two Commonly Used GPS Devices and Its Possible Application for Their Management and Conservation

Giraffes taught a computer to recognise drinking without seeing water. Lowering the head to drink creates a distinctive movement pattern, quite unlike reaching up for leaves. If those patterns can be recognised automatically, researchers could learn what a giraffe is doing even when they cannot follow it closely enough to watch.

Stefanie Brandes, Florian Sicks and Anne Berger fitted head-mounted movement sensors to three trained giraffes in Canadian and German zoos. They recorded the animals’ behaviour alongside the sensor readings, then used these matched examples to train computer models. Two devices were compared, and a crucial test asked whether a model trained on some giraffes could classify the movements of another.

Drinking and feeding above eye level were among the easier activities to recognise in the within-animal analyses. Standing and chewing cud were more easily confused with other behaviours, because similar head positions can serve different activities. Performance fell when the model encountered a giraffe excluded from training. Only three captive animals were studied, and recordings combining several behaviours were omitted, so success on wild giraffes was not established. The promise is specific: head movement contains recognisable clues to daily life, but interpreting them requires learning both shared patterns and individual differences.