10.5061/DRYAD.0RXWDBS2R
Resheff, Yehezkel
0000-0001-7863-7632
Holon Institute of Technology
Bensch, Hanna
0000-0002-8449-9843
Linnaeus University
Zottl, Markus
Linnaeus University
Rotics, Shay
University of Cambridge
Data from: Correcting a bias in the computation of behavioral time budgets
that are based on supervised learning
Dryad
dataset
2022
body-acceleration
bio-logging
behavioral time budget
biotelemetry
animal behaviour
FOS: Biological sciences
European Research Council
https://ror.org/0472cxd90
742808
Crafoord Foundation
https://ror.org/02hwwbr17
2018-2259
Crafoord Foundation
https://ror.org/02hwwbr17
2020-0976
2022-03-31T00:00:00Z
2022-03-31T00:00:00Z
en
8253012 bytes
8
CC0 1.0 Universal (CC0 1.0) Public Domain Dedication
Supervised learning of behavioral modes from body-acceleration data has
become a widely used research tool in Behavioral Ecology over the past
decade. One of the primary usages of this tool is to estimate behavioral
time budgets from the distribution of behaviors as predicted by the model.
These serve as the key parameters to test predictions about the variation
in animal behavior. In this paper we show that the widespread computation
of behavioral time budgets is biased, due to ignoring the classification
model confusion probabilities. Next, we introduce the confusion matrix
correction for time budgets -- a simple correction method for adjusting
the computed time budgets based on the model’s confusion matrix. Finally,
we show that the proposed correction is able to eliminate the bias, both
theoretically and empirically in a series of data simulations on body
acceleration data of a fossorial rodent species (Damaraland mole-rat,
Fukomys damarensis). Our paper provides a simple implementation of the
confusion matrix correction for time budgets, and we encourage researchers
to use it to improve accuracy of behavioral time budget calculations.
We obtained this dataset from 16 Damaraland mole-rats (DMRs) that were
collared with acceleration loggers (Technosmart LTD, Italy) for 1-3 weeks,
and videotaped during this period to match the acceleration records with
known behaviours.
See README.txt