Limitations of using surrogates for behaviour classification of accelerometer data: refining methods using random forest models in Caprids

Nikki Marks,Joshua P. Twining,Philip A. Stephens,Eleanor R. Dickinson,Jennie Westander,Rory Wilson,David M. Scantlebury

doi:10.1186/s40462-021-00265-7

Abstract

BackgroundAnimal-attached devices can be used on cryptic species to measure their movement and behaviour, enabling unprecedented insights into fundamental aspects of animal ecology and behaviour. However, direct observations of subjects are often still necessary to translate biologging data accurately into meaningful behaviours. As many elusive species cannot easily be observed in the wild, captive or domestic surrogates are typically used to calibrate data from devices. However, the utility of this approach remains equivocal.MethodsHere, we assess the validity of using captive conspecifics, and phylogenetically-similar domesticated counterparts (surrogate species) for calibrating behaviour classification. Tri-axial accelerometers and tri-axial magnetometers were used with behavioural observations to build random forest models to predict the behaviours. We applied these methods using captive Alpine ibex (Capra ibex) and a domestic counterpart, pygmy goats (Capra aegagrus hircus), to predict the behaviour including terrain slope for locomotion behaviours of captive Alpine ibex.ResultsBehavioural classification of captive Alpine ibex and domestic pygmy goats was highly accurate (> 98%). Model performance was reduced when using data split per individual, i.e., classifying behaviour of individuals not used to train models (mean ± sd = 56.1 ± 11%). Behavioural classifications using domestic counterparts, i.e., pygmy goat observations to predict ibex behaviour, however, were not sufficient to predict all behaviours of a phylogenetically similar species accurately (> 55%).ConclusionsWe demonstrate methods to refine the use of random forest models to classify behaviours of both captive and free-living animal species. We suggest there are two main reasons for reduced accuracy when using a domestic counterpart to predict the behaviour of a wild species in captivity; domestication leading to morphological differences and the terrain of the environment in which the animals were observed. We also identify limitations when behaviour is predicted in individuals that are not used to train models. Our results demonstrate that biologging device calibration needs to be conducted using: (i) with similar conspecifics, and (ii) in an area where they can perform behaviours on terrain that reflects that of species in the wild.

Highlights

Biologging has transformed what we know about wild animal behaviour [1,2,3], with particular value attributed to tri-axial body acceleration [4,5,6]
We provide a widely applicable template for refining the use of random forest models to predict behaviours including; feature selection approaches, the addition of tri-axial magnetometry variables, selecting the optimum sampling frequency, handling unbalanced observations and data splitting method
Seven variables were removed due to them being highly correlated and a further 13 variables were removed in recursive feature elimination (RFE), with 17 variables included in the final model (Fig. 2; Additional file 2 Fig. S4)

Summary

Introduction

Biologging has transformed what we know about wild animal behaviour [1,2,3], with particular value attributed to tri-axial body acceleration [4,5,6]. Where data are limited by direct observations [9] or telemetry is constrained (e.g. sampling intervals are low [10], location is inaccurate [11, 12]), these devices record body movement of animals at high frequencies. They can provide detailed information on the study subjects, representing a powerful opportunity to study enigmatic species [6]. Random forest models are a commonly used approach for classification of behaviours from accelerometry data and provide high accuracy [4, 18].

Methods

Results

Discussion

Conclusion